Saturday, October 31, 2020

 Soil Permittivity under Antennas

There is a lot of speculation about the electrical properties, at RF frequencies, of the soil under antennas.  Thankfully, there's an interesting data set where some probes were buried near one of the antennas at the Owens Valley Radio Observatory Long Wavelength Array (OVRO-LWA).  OVRO-LWA is an array of hundreds of dual polarization fan dipoles that is used as a radio telescope over the 27-85 MHz frequency range.

The probes measure the properties at 50 MHz, and produce data both in terms of the complex permittivity and in the more familiar relative permittivity (Epsilon) and conductivity (sigma, in Siemen/meters). The data were taken over about a 6 month time span including the wet and dry seasons.

As described in the data page, the sensors were buried 4", 14" and 21" deep next to one of the antenna stands. From the page:

 "The test pit was dug down to hardpan, which in this location is 22" below ground level. The strata is clean sand down to ~18". From there, the fraction of a green clay-like material increases with depth, reach hardpan as noted above.

The anticipated depth of the hardpan, based on Andres Rizo's experience at site was 10-20'. The actual depth at any given location will depend on river course wander over millenia and the presence of standing surface water."


Data from https://github.com/ovro-lwa/leda_docs/wiki/Soil-Permittivity-Sensors

The actual sensors https://stevenswater.com/resources/datasheets/HydraProbe%202020_draft3.pdf


And, now, on to the data: Here's all of it on one plot - the top sandy layers have low epsilon and low conductivity. The deep clay layer shows distinct clumps, related to the soil moisture. Spring is the wet season in the Owen's Valley of California (just east of the Sierra Nevada), both from rains, and from snow melt, although I don't think there's any melt-fed streams in the OVRO-LWA area.  This plot is in complex permittivity.  

All of the data
Plotted as real vs imaginary epsilon

21" deep probe

Here's the clay layer plots - first plot is complex permittivity as e' and e''. Second plot is in the more familiar epsilon/sigma terms.  Epsilon = e' + j e''; Sigma = 1/(omega e'' * e0)

  
 

14" deep probe

Moving up to the sand, you can see it varies a bit more, and more abruptly.



4" deep probe

And for the top layer of sand, it varies all over the place as the soil is damp and dries.





Take home

There are significant variations in epsilon with soil moisture content. The clay (21" measurements) is significantly higher conductivity, even during the dry season.

If you're modeling your antenna using a tool like NEC, you should model with several combinations of epsilon and sigma, and make sure you understand the performance differences.  Using the 13/.008 "good soil" may not be very representative of YOUR location. And particularly for verticals, the soil properties in the far field makes a big difference.