On Sun, 02 Nov 2003 03:44:03 -0800, Roy Lewallen
wrote:

|I haven't had the time to participate in this, but in a brief look, it
|looks pretty silly. Of course EZNEC shows no current difference across a
|load. The EZNEC model of a load has zero length, so the current at both
|terminals has to be the same. You will see a current change over the
|length of a model of a conductor, because it does have length. The coil
|in the web site pictures certainly has length, so why should you be
|surprised to find a current change over its length? Did the experimenter
|perhaps do the same test with the meters placed the same distance apart
|with just a conductor in between? Would there be some great revelation
|in finding that the current was different at the two points?

I too have been lurking and while I didn't spend any time reading the
eham stuff what was going on here got me looking at a model of this
situation.

I used MultiNEC to invoke EZNEC for all calculations. I modeled a
shorter-than-quarter-wavelength vertical, loaded with an inductor, all
of this over perfect ground. Using MultiNEC, I used equations to
change the length of the radiator, the position of the inductor, keep
the segment length as short as guidelines allow and resonate the
result after each change.

Nothing I did solved this argument but I did make a couple of slightly
related observations. Unless I'm mistaken (always a distinct
possibility):

1) When the radiator is electrically very short and near resonant the
current does not follow the classic patterns shown in most of the ham
literature, i.e. nearly constant below the inductor and close to a
straight line taper above. The current actually peaks at the inductor;
in other words, the highest current point on the structure is at the
inductor.

Hanson's paper (Robert C. Hanson, "Optimum Inductive Loading of Short
Whip Antennas", IEEE Transactions On Vehicular Technology, Vol VT-24,
No. 2, May 1975, pp 21-29) shows this, although his graphs show a
steady decrease in current from the feedpoint to where the peaking
begins. I did not see that, but instead a steadly increasing current
from the feedpoint to the current peak.

Not only that, the peaking is almost independent of inductor Q.
"Almost" meaning that my model shows that the current is actually
slightly higher in a lower Q inductor.

2) The structure Q, defined as the change in reactance with respect
to frequency, is independent of inductor Q.

3) For a give length radiator, gain is unaffected by where the
inductor is located along the length of the radiator and by inductor
Q.

If 3 is correct then I can remove the inductor from the radiator
without effecting the gain and place it before the feedpoint to
resonate the structure. Once out of the radiator, the current through
the inductor is constant. Just like it is in the antenna if it has
zero length.

By now you're all saying, "Huh?" But remember, this is for an antenna
over perfect (zero loss) ground.

So instead of worrying about perfecting our antennas, we should be
trying to perfect ground and/or zero length loading coils.

|
|I was intrigued by the claim that a toroid measured significantly
|different from one end to the other. I wonder if the tester tried
|reversing the meters to verify that he got the same reading in both
|cases. If he did, I'd be interested in learning more details.

Considering that *anything* inserted in the structure upsets the
current distribution, as Roy says above, why wouldn't the answer be
different. Even the toroid, or the distance between the insertion
points, have *some* length.
|
|Unfortunately, the main objective of the web site seems to be to insult
|Tom, W8JI, rather than to be objective. So in my mind that leaves the
|possibility open that the experimenter is more interested in finding
|evidence that would disprove Tom than in presenting carefully measured
|and objective data.

Eureka!


Wes Stewart N7WS