[Top] [All Lists]

[Towertalk] Models and real world --oops

To: <>
Subject: [Towertalk] Models and real world --oops
From: (Tom Rauch)
Date: Thu, 11 Jul 2002 09:00:58 -0400
> At 09:03 AM 7/10/02 -0400, Pete Smith wrote:
> >it can be useful to have these numbers, even though we know the
> >experimental accuracy may be in the 1 dB range.
> For some reason, the reflector strips out the mathematical symbols for
> plus or minus.  That's what I meant, "plus or minus 1 dB."

Hi Pete and all,

My point is best illustrated by the A-B comparisons of Rick, N6RK, by 
comments from I4JMY, and by my own experiences.

Rick (as I have) has found the losses in verticals not as bad as 
predicted in actual tests, and documented what appear to be 
discrepancies in ground effects. I'm sure there are probably cases 
going the other way.

The problem is, we give exact answers based on models (and 
measurements) that are never exact. We often model antennas and 
assume they are within a dB or two of the real world, but I'd wager 
that is not true in many if not most cases. I'd also bet many of us 
even assume they are so perfect even the tenths of a dB count!
I'm sure the models are very accurate if the antenna is far from 
earth (in wavelengths) or over a large conducting reflector, and if 
nothing goes outside the limits of the model. But in the case of a 
simple vertical or an array of verticals near earth (within 1/4 wl or 
so of earth) I would expect discrepancies that could sometimes be 
significant to making decisions.

Worse yet we have a tendency to compare system hundreds of miles 
apart with different equipment, locations, and operators to decide 
what antennas "work well" and what antennas "don't work".

I am concerned an over-reliance on computer models that we assume are 
"absolute" has collectively given us the impression that we can 
predict the real world, and we are as a group losing touch with 
understanding how to make valid comparisons between systems. 

The same changes are creeping into the ionospheric soup, where we see 
silly things like 20-dB coupling loss that can't be measured.

Over reliance on models and under reliance on common sense or a feel 
for what is physically happening in a system doesn't exactly push us 
towards greater knowledge. A lot of the things we now assume are 
concrete numbers are really nothing but best-case trends with many 

73, Tom W8JI 

<Prev in Thread] Current Thread [Next in Thread>