Cecil Moore wrote:
Keith Dysart wrote:
You have done this before; postulating
explanations that only work in the complexity
of the "real" world, but fail when presented with
the simplicity of ideal test cases.

For Pete's sake, Keith, Ohm's law doesn't even
work when R=0.

Then, when the explanations fail on the simple
cases, claiming these cases are not of interest
because the real world is more complex.

I define the boundary conditions within which my
ideas work. Whether they work outside those defined
conditions is irrelevant. I believe they do work
for ideal conditions, but I don't have the need
to prove a "theory of everything".

Every model that we use has flaws. Asking me to
come up with a flawless "theory of everything"
model is an obvious, ridiculous diversion but
you already know that.

This isn't a diversion: it's the core of the whole dispute.

These days, mathematical models are the normal, everyday way that
engineers go about their business. A bedrock principle is that if a
model is going to be usable and trustworthy, it MUST join up correctly
with existing knowledge. Your model can be as elaborate as you like, but
it always has to prove itself against the simple cases that we already
know about.

Anyone with experience knows that these "simple" reality tests are the
most often the hardest for an elaborate model to pass... but that
doesn't excuse them from the test. If a model cannot handle the simple
situations that we do understand, we can never trust it in more complex
situations.

Ohm's law is a perfect example of a model that works. The whole point is
that Ohms' law IS a good model of reality for a very wide range of
situations, including the simple but extreme case where R equals
exactly zero. It's absurd to suggest that there's a glitch - it simply
means that V would be exactly zero too.

Likewise there are no glitches in the standard circuit models for
inductance and capacitance. They work just fine, for all cases where the
dimensions of the circuit are very small with respect to the wavelength,
so that distributed effects and radiation are negligible. Where those
assumptions are no longer accurate, we can extend the simple model to
include some corrections. But the most important point is, we always
know that we're building up from a solid foundation.

That is also the sensible way to think about loaded antennas. Calculate
it the simple way first, assuming lumped inductive loading, and then
apply corrections as necessary. As I've said before, this simple, solid
method is the one that works. It can take you straight to a workable
prototype, which can be quickly adjusted to frequency. Countless authors
have demonstrated how to do this, and anyone can download G4FGQ's
MIDLOAD program to do the same.

While other people choose to build on those solid foundations, Cecil
insists that simple routine reality tests are a "diversion". He prefers
to keep his floating castles well clear of such hard rocks.


--

73 from Ian GM3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek