LA4RT Jon wrote in :

....
N6LF made extensive measurements and essenstially confirmed this. He
wrote a 7 part series of articles for QEX. You can download them at
his site:
http://www.antennasbyn6lf.com/2009/1...les-on-ground-
system-experiments.html

Hi Jon,

All of the articles are interesting. I have previously read the last, but
wasn't aware it was only one of a series.

Looking at Article 3, Fig 1 suggests that efficiency improves very slowly
beyond about 32 radials lying on the surface, and 4 such radials are
about 5.4dB below 32 radials.

In Fig 2, he shows 4 radials just 6" (150mm) above ground as about 5dB
better than 4 radials on the ground.

My interpretation of Fig 1 and Fig 2 then is that 4 radials at 6" are
about 0.5dB behind 32 radials lying on the ground. That is the type of
effect I was referring to when I said "If my models are valid, and they
use 'average ground', the indication is that while it may require a large
number of buried radials (16) before efficiency levels off a bit,
similar efficiency can be obtained with just three radials elevated more
than 100mm above the soil" in my first post.

Rudy reports some further small improvement (1dB) in raising the radials
to 4'(1.2m).

These are very similar effect to those predicted by my NEC4 model. In the
case of my model of radials from 100mm depth to some distance above
ground, the improvement was mostly in the range of heights from 0mm to
about 20mm. Obviously, the model is sensitive to soil type, and different
soil types can be expected to yield different response... but it would
seem that just 3 radials at 1 to 3m height give similar system efficiency
to 16+ radials shallow buried for a range of common soil types.

I know my interpretation of Rudy's measurements and my NEC4 model don't
fall in line with some traditional thinking, and will not appeal to many.

I think it is time to build a prototype.

Thanks for the links.

Owen