----- Original Message -----
From: "Jim Lux" <jimlux@earthlink.net>
> In the cable TV world, they tend to use dBmV (into 75 ohms, presumably),
> but in any case losses should work the same regardless, a typical path
from
> "mainline cable" appears to be something like 26 dB for the tap, 2dB for
> the drop, 6dB for inhouse cable and splitters, so 34dB from "main wire" to
> the -40 odd at the set.(I seem to recall something like -33 dBm
corresponds
> to +15dBmV, which is a typical level). Figure at least zero, and then
> they'd add some margin, so they're probably running +10 dBm (per carrier)
> or several watts down the wire with all 100+ channels...
Generally I think the carrier levels on analog coaxial trunks are
set at something like +33dBmV/carrier (~-15dBm). You can't run
the carriers much higher on the trunks or the cumulative IMD and
crossmod will kill your system performance. That's why you have
trunk amps every 1500 to 2000 ft. Branches from the main trunk
(line extender amps) are typically run a bit hotter: +44 to
+48dBmV/carrier (-4 to 0 dBm) because you no longer have to
worry about cascading N amplifiers when you reach the end-of-the-line.
Of course now many systems are hybrid fiber-coax (HFC). I haven't
a clue how they set the levels in those systems.
> Say you've got +10, and you've attenuated it 100 with the shield, you're
> down to -90, which is still easily above the noise floor with a narrow
band
> receiver (like an aircraft or public safety radio) (I think that there's a
> hole in the cable channel band plan to avoid the aviation bands, just for
> this reason?)
>
> The FCC spec is 20 microvolts/meter 3 meters from the point of leakage. I
> think that's about -90dBm into a dipole. There's some space loss over the
> 3 meters (but not a huge amount maybe 30-40 dB), so if your signal leakage
> starts getting up into the -50dBm or -40dBm range, you'd have a problem.
>
The main problem with the RFI numbers thrown around in the
CATV industry is that they aren't well defined. The guy I used
to work for had a large galvanized steel drum (~55 gallon size)
with a hardline protruding down thru the lid along the drum axis
with an F fitting on it. Surrounding this hardline was ~6" diameter
cyclinderical galavanized duct that was insulated from the outer
body of the drum. This 6" duct was supposed to represent the
center conductor of a very large coaxial structure. The output of
this EMI chamber was a connection between an F-connector
on the side of the drum and the 6 inch diameter center
conductor.
We would place test samples (usually CATV splitters) on the
end of the hardline protrusion and then slip the lid over the drum.
We would then drive the hardline with a tracking generator and
"sniff" on the output port of the drum with a spectrum analyzer.
The drum was supposed to mimic concentric coaxial lines, but
in reality it was way overmoded at VHF/UHF frequencies. It was
essentially a big resonant cavity. As I recall the peak response
was around 400 MHz. Everywhere else, EMI performance of the
test samples looked really good because of the cavity's intrinisic
selectivity. At the resonant frequency, however, you could get a
gross feel for how bad a component leaked. If you amped up the
tracking generator output to +25dBm and narrowed the resolution
bandwidth of the spec A, you could establish a dynamic range
of about 120dB. Sure enough, if you put a little 400 MHz monopole
in the chamber you would get a big response (something like -5dBc
relative to the tracking generator thru cal). Then if you installed a
shielding 75 ohm termination for your test article and got the
connector threads real tight, the spectrum analyzer reading would
drop to the circa -120dB level. Solder sealed splitters with claimed
RFI of -130dB would be about as good as the shielded termination.
Roll sealed passives with claimed RFI of -100dB would come in
around -80dB in many cases.
This perverted EMI chamber apparently was based on a paper
that some guy wrote up in one of the CATV trade journals. It was
my understanding that it was the defacto measurement standard
for RFI specs in the CATV business. The origins of the chamber
appear to be some work done at Belden whereby Belden engineers
would suspend a long span of coax inside a long concentric coaxial
structure and then measure the cross talk between the inner coaxial
structure (line under test) and the outer coaxial structure (test
chamber). Of course, unlike the pregnant CATV version, the Belden
instrument was long and slender, properly terminated, and operated
below waveguide mode cutoff :):) In either case, these shielding
effectiveness numbers used to bother me because it was never clear
how to relate them to absolute leakage levels. I always just assumed
that -100dB was relative to an istropic radiator, but even that would
be bogus for a cable because the amount of leakage for a cable
with -100dB shielding effectiveness would surely depend on its
length (e.g. you would expect 2 inches of -100dB cable to leak a
lot less energy than 200 ft of -100dB cable). At most the numbers
are relative figures of merit (e.g. -120dB cable is 20dB better than
the same length of -100dB cable). As to whether one needs
-120dB cable or -100dB cable in any given situation is probably
best determined experimentally :):)
In any case, after using that drum chamber for a while, it became
very clear to me how important connectors are to the equation. The
difference between a loose and a tight F-connector can be many
dB.
73 de Mike, W4EF...............................................
_______________________________________________
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|