NEC-LIST: Modeling and Measuring

Roland:

Thanks for your email. If I gave the impression that I thought that
because we had better model tools now with the method-of-moments, etc.
and no longer needed measurements then I certainly stand corrected. I
am a major supporter of measurements, but it is cost-prohibitive, and
sometimes time-prohibitive, to rely totally on measurements. I worked
with LLNL (Jim Breakall) and NPS (Dick Adler) on checking out NEC-3 at
the Livermore, CA campus. I also took measurements of an antenna (a
Spiracone) on the Stanford University campus for Jim Breakall when he
was still at LLNL. I also used the SRI OWL Kit for measuring the soil
conductivity and relative permittivity at each site vs frequency to
provide input data for the NEC-3 calculations. Jim got negative input
resistances for the Spiracone over part of his range, but eventually
found the problem and got something like my measured antenna input
impedances after working the problem with two grad students at Penn
State. That attests to your comment about the problem of getting the
modeling right for real-world problems without also having some good
measurements to check it out. In June 1974, I gave a paper at the IEEE
EMC Symposium in New York City on the subject of "On the Interplay
Between Analysis of Radio System Performance and Measurements of
Man-Made Radio Noise." At the time, I was concerned about system
performance predictions that didnt work because of unrealistic noise
assumptions (usually assuming too little noise). I eventually "got it
right" in modeling the operational range of a radio system in the
Sept. 1980 IEEE Trans. COM, but I had to go and measure antenna gain
on a helicopter in order to get my model to match reality. In that
case, the antenna gain in the band 30-40 MHz for a "tail fin" antenna
had been given too optomistically in some published materials. My
system performance model in that case predicted the probability of
achieving a certain intelligibility score as a function of range, with
the operational range being that range when the probability had
dropped to a user-specified required range.

In the case of the Wheeler can method, I was suggesting using the can
to get an approximate measure of the loss resistance, Rl. I was
suggesting using a MOM model to get an estimate of Rr, the radiation
resistance. Since the input impedance of the antenna (Ra) could be
measured as well as predicted, one could check the measured value of
Ra with the MOM-predicted value of Ra, and with the sum of the can Rl
and the MOM Rr. The model provided a way to get the Rl where before
we needed to do something like the Wheeler can method to estimate Rl.

So I believe we agree on the importance of measurements and of the
interplay of measurements and modeling and analysis.

Sincerely,

George
Received on Thu Apr 22 1999 - 18:46:41 EDT