On 3/16/18 12:33 PM, Jim Brown wrote:
Here's measured data for Harbor Industries RG400.
k9yc.com/RG400DataMF-HF.png
Data was obtained by S11 sweeps (1-30 MHz) using VNWA, exporting as S1P
files to AC6LA's Zplots spreadsheet for Excel. ZPlots does the math to
compute and plot Zo, VF, and attenuation vs frequency, and finds the
constants K0, K1, and K2 in the Johnson transmission line model. The
lower plot is computed from the raw data, the top two plots come from
the model. Zo (nom) and VF (nom) are the values to which the curves are
asymptotic at VHF.
These curve shapes are predicted by the wave equations and are typical
of virtually all transmission lines.
Interesting data..
So I wonder what truly horrible coax would look like - that is Z vs freq
or loss vs freq are generally smooth curves (and, so, well modeled by
k0,1,2, etc.)
I wonder if a "figure of merit" for coax could be developed by
"deviation from model" - that is, we all know RG-174 has more loss than
RG-213 or 1" hardline, but that's mostly a matter of geometry and materials.
But maybe "good" coax has a better fit to the model than "poor" coax -
or the plots move around as you move the coax or there's perodic dips
and peaks (at an enlarged scale) from a manufacturing roller that has a
flat spot, or the cable being stored with a big weight on the spool on
one side, or....
The CATV business has measurements and criteria that attempt to
characterize shielding effectiveness (not in an idealized modeled sense,
but as in "what you measure in a standard test fixture"), and I'm sure
there's some way to measure IMD from cable/connectors.
Is there some "doable by a ham with simple test equipment (which today
means an inexpensive 1/2 port VNA)" test that can distinguish between
"good" and "less good" coax?
Preferably with a one port measurement (so you don't have to have access
to the inside of the spool or other end of the cable).
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|