Well, I guess I didn't really learn as much as I thought I did. I took
your advice and modeled a 450 MHz dipole instead and got a curious
result. I split the dipole into 31 wires each 2mm thick and 10 mm long,
with the feed in the center (wire #16). I added insulation (Er = 2,
thickness = 5mm) as balanced pairs, as in #1/#31, #2/#30, #3,#29, etc.
I found I was getting slightly squirrely results if I only did one side
of the dipole at a time ... I think the imbalance was swinging the
result a bit.
Here are the EZNEC Pro/2 results for resonant frequency. I'm reading
the numbers off the SWR plot so the accuracy is not better than +/1 at best.
no dielectric loading = 453 MHz
1/31 = 395 MHz
2/30 = 448
3/29 = 449
4/28 = 450
5/27 = 449
6/26 = 448
7/25 = 447
8/24 = 446
9/23 = 446
10/22 = 445
11/21 = 444
12/20 = 443
13/19 = 443
14/18 = 442
15/17 = 441
16 = 453
The end sections clearly show the greatest effect and it is a loading
effect that decreases the resonant frequency. This is what I originally
hypothesized ... that the dielectric loading effect would occur where
the E-field is highest. And at first the resonant frequency increases
as I move the dielectric loading toward the center ... as expected.
But to my great surprise, as soon as I get a little further away from
the ends (at roughly wire #4 going toward the center) some other effect
takes over that DECREASES the frequency going toward the center, but
very gradually and starting from a frequency almost equal to having no
loading at all (448 MHz versus 453 MHz). And then at the center wire
(where there is very little E-field) the resonant frequency jumps back
up to the same as if there was no dielectric loading at all. I've
double checked the runs and right or wrong that's what they are.
The current in a half wave dipole follows a half-sine wave profile, but
the voltage follows a more extreme cotangent-like profile. The voltage
increases VERY quickly in a very short distance at the ends of the
dipole, and my model was done with lossless wires so there really isn't
a limit to it. I'm using a 1 amp forcing function for the source at the
center of the dipole, EZNEC says that the current at the very end
segment of the very end wire is 0.04 amps. If I did the math properly
that works out to be over 30,000 volts per the model. The next 10mm
wire closer to the center (wire #2) is already down to about 1,000 volts.
I should probably run the numbers over again with some resistance in the
wires, but I think that would only moderate the magnitude of the effect
... not the general shape of the profile.
Going back to the full wave loop I described earlier, I didn't see this
same kind of result because that model didn't have the same
granularity. In that model have 40 cm long wires (7 of them split into
17 segments) on each side of the loop, and when I look at the current
distribution on the center wire (#6 in that model) it is clear that the
insulation for that wire spans enough of the wire that the loading
effect of the high E-Field in the center segment of wire #6 is swamped
out by the lesser effect of the adjacent segments. Instead, I only saw
the secondary effect similar to what is shown by dipole wires #5 though
#15 above.
I think I understand WHAT I'm seeing now and the end effect makes sense
to me, but I still don't understand what is happening as we move toward
the center of the dipole.
Or maybe EZNEC simply doesn't handle all of that very well and I'm
letting myself be led around by the nose.
Dave AB7E
p.s. At my age (almost 78) understanding something isn't necessarily
the most important part, although it's frustrating if I don't. Thinking
about it, as in keeping my mind active, is probably more valuable. ;)
On Tue, 28 Jan 2025 12:08:57 -0700, David Gilbert <ab7echo@gmail.com>
wrote:
I did some more searching and found a similar discussion on the 4NEC2
reflector from about 3 years ago. In it was a link to this article by
Cebic which seems to indicate that dielectric loading of a conductor
produces an inductive effect. If that's the case it makes sense that it
would have more effect at a current maximum than a voltage maximum.
https://hamwaves.com/wire/doc/Insulated%20Wires:%20The%20NEC-2%20Way.html
I also harkened back to the formula for velocity factor in a coaxial
transmission line, which is V = 1/sqrt(er), where er = the relative
permittivity. That certainly describes the equivalent of an inductive
effect.
Apparently at age 77 I'm not yet too old to learn ... although that day
might not be too far off. ;)
Dave AB7E
>
>
> On Mon, 27 Jan 2025 20:33:55 -0700, David Gilbert wrote:
>
> I've been using EZNEC Pro/2 to try to model the effects of dielectric
> loading on a wire antenna ... specifically a full wave loop on 435 MHz.
> I split the sides of the loop up into multiple wires (each with multiple
> segments) so I could individually declare a thick dielectric ... i.e.,
> insulation in the wires table ... separately for each portion of the
> loop. To my surprise the loading effect seems to be greatest at or near
> current maximums, not at voltage maximums where I would have presumed
> the electric field would have the most effect. This has me greatly
puzzled.
>
> Any thoughts?
>
> Dave AB7E
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|