Hi Roy,

Thanks. I went back and bravely ran the original coil model, ignoring
the guideline checks. The difference in results from the two coil
models, eight-sided and six-sided, was in the noise. (Or at least well
below the new RRAA error guideline of 59%.)

Emboldened by this apparent success I tried to substitute the now-famous
W8JI coil; 100 turns, 2 inch diameter, 10 inch length. This time the
overall 500 segment count was the limiter (I am cheap), so I had to make
the coil four-sided. The delay through this coil was about 9 degrees at
5.89 MHz and about 6 degrees at 3.9 MHz.

Sooooo, the bottom line for the 17,000 posts in this thread is that the
pure lumped-circuit toroidal coil analog might be just a slight bit
stingy, but it comes pretty close to reality. The full-bore transmission
line model for the coil works as well, but it does not appear to add
much useful information. The real phase shifts are well below the 15
degree transition point claimed by the Tesla coil crowd. I am sure none
of this is news for you or for most others.

Cecil has thanked me on several occasions for bringing up the standing
wave equations. I now must return the favor by thanking him for setting
up the traveling wave model. It did not prove his point, but it helped
bring clarity to the issue.

I am confident that this will not be the end of this immortal thread,
but it is now quite clear through simulation that the entire issue is
much ado about very little.


73,
Gene
W4SZ


Roy Lewallen wrote:
Gene Fuller wrote:

. . .
I tried to adjust the frequency downward, but the segment length limit
is reached for the coil model. I redrew the coil with only six
segments per turn, rather than eight. Now the frequency can be lowered
to about 3.94 MHz without EZNEC guideline check warnings.
. . .


Of the Guideline Check warnings, the one about short segment length can
most often be ignored without a problem. If in doubt and where
practical, do a test case. For example, model the inductor with a wire
through the middle directly connecting top and bottom and with a source
in the middle. Note how the source reactance changes with frequency to
see that it follows what it should theoretically do. Another test you
should run when in doubt is the Average Gain test, described in the
manual. Either is probably adequate to have confidence in the results.

The Guideline Check warnings are based on NEC-2 recommendations. I've
found that you can often make segments much shorter than the
recommendations without causing a problem.

Roy Lewallen, W7EL