On 10/27/2011 1:04 PM, Owen Duffy wrote:
Jim wrote in
:

On 10/26/2011 12:37 AM, Owen Duffy wrote:
Jim wrote in news:j846r5$7h3$1
@news.jpl.nasa.gov:

There's a goodly bit of empiricism in trap design.

And a goodly bit of misinformation in some of the traditional ham
sources.

I have had an interest in the so called coax traps, and just in the
last week or so, make some headway with some good measurments of the
underlying inductor formed by the coil of coax shield. There were
some surprises. For those interested, see
http://www.vk1od.net/antenna/coaxtrap/index.htm .

Short of performing Medhurst style measurements on a range of
inductors, there is no way to be sure that the effects observe apply
generally, or what a more general model might be.


There is a fair amount of literature on L and C (and loss) for
shielded inductors, which is what a lot of traps look like.. the ones
which use the C to the surrounding tube, anyway, like in the
4,5,6-BTV. Some papers have generic cookbook-ey design equations
which might be useful (although I don't have any citations off the top
of my head).

For a bare coil of the appropriate L/D ratio, Medhurst will get you in
the right starting place.

The problem would be things like manufacturing variability, if you're
copying (or writing instructions) one-off design. Stuff like 'how
thick is the enamel/polyurethane/PVC insulation'.

Sort of like the measurements of Z and loss for zipcord. The
dielectric properties aren't controlled in manufacturing, so what you
measure on brand X, 16 Feb 2001 may have little or no relation to what
you measure on brand X, 20 Oct 2010 vintage.

All noted.

Many readers will recall my interest over more than a decade in
predicting the effective RF resistance of the outside surface of a
braided coax shield, especially when it forms a solenoid... as in the
coax traps.

The last round of measurements by VK2KRB were most interesting, because
they strongly suggested that Q was not proportional to root of f as
Medhurst and predecessors observed for round copper wire (and of course,
R was higher than for an equivalent sized round copper conductor).

Jim, interesting that you mention ZIP cord. I have seen a number of
articles recently discussing the TL characteristics of ZIP cord, and
again many readers will recall Jack Smith's measurements published here
about 10 years ago.

I recently put a new TL calculator up, it uses input parameters of Ro,
vf, k1, k2 to solve problems similarly to the older TLLC. It is at
http://www.vk1od.net/calc/tl/atllc.htm . I wrote an article with some
examples of using it at http://www.vk1od.net/calc/tl/atllcEx.htm .
Example 2 plots Jack's data on ZIP attenuation from back then. A
stunning set of measurements, and statistically tighter than I have seen
from any one else. Nevertheless, I see a wavelike shape to the error
between actual and the model, a growing sinusoid that prompts the
question of why, was it some common mode effect and radiation.


Interesting..
So the loss is about 0.12 dB/meter at 10 MHz, compared to, say, RG-58 at
0.04 dB/meter. That's a huge difference, especially since the center
conductor on the RG-58 is probably smaller than the zip cord (what was
the wire size? I think you cited 1.22 mm diameter? AWG 16? or AWG 18?)