"Reg Edwards" wrote in message
...

In the case of RADIAL_3 the obvious purpose of the program is to
assist with choosing an economic length and number of radials to be
used with a given test antenna height. It is also educational in that
after reading the introductory notes and using it, the user will have
a better understanding of how radials work.


I understand that Reg, but somewhere in all this an important consideration
is being lost. The issue isn't +/- some questionable percentage of accuracy,
it is the underlying assumption in the model you are using to arrive at the
"much shorter than everyone else's radial length". You have to admit,
predicting 90% efficiency with 5 metre long radials (26 in my example) is
stunning...given BL&E, Tom's measurements, and the yet to be run NEC-4
analysis. We aren't talking small differences here, we are talking NIGHT and
DAY in terms of length.

It really is this simple: Your program predicts neglible current at
distances greater than 5 metres in the example being discussed. Your program
says that any further lengthing is borderline useless. Tom's measurements
completely disagree. BL&E, I am told (I haven't read that section) also
completely disagree. I'm waiting to see what NEC-4 says.

They key is this: are their ANY soil conditions wherein your model of 5
metre long radials (26 of them) will agree with the existing experimental
data, or NEC-4 modeled data? If not, then the "radial as transmission line"
model fails, and should not be used. If one doesn't get the 20 to 25 dB of
attenuation within the radial length limits your program predicts, then the
program is in error and will lead to false conclusions...not just
"inaccuracies", outright major errors. I would love to put in 66 radials 5
metres long and know that they work every bit as well as 66 radials 18
metres long. It would save a lot of money in copper and extra lawn staples.
If, on the other hand, your model is wrong, then a lot of work has been done
for next to nothing. If the purpose of the program is to help in this
process, the program must be trustworthy in its MAJOR assertions.

================================================== =====
We need to know: does the predicted attenuation of current along a radial
wire happen as quickly as you predict? This can be measured. This can be
modeled. That's what makes this fun. Let's find out. Let's see what agrees
with what and what doesn't. Then we can conjecture as to why, and which
approach is to be "believed".
================================================== ======

I'm not denigrating your work. I have all your programs and play with many.
I have found several to be wonderfully useful. However, when something is
called into question, I'm just not religious in scientific matters. Even the
Qur'an says, "Bring your proof, if ye are truthful." (sorry, I couldn't
resist)

73,

....hasan, N0AN