Hi Jack,
Coaxial connectors leak, but generally not very much. The amount depends
on the quality of construction, the applied torque, and the frequency. I
worked in the CATV industry for a few years in the early 90's and I
remember cumulative leakage index (CLI) being a big deal. The operating
frequencies of Cable TV systems overlap the VHF/UHF ham-bands, public
service communication frequencies, over-the-air television frequencies,
and aeronautical frequencies. To successfully coexist with these
services, CATV plants have to be well-shielded.
Customer premises equipment (CPE) such as F-connectors were a source of
trouble as were passive devices (e.g. RF splitters and couplers). The
really low-end passives were housed in zinc castings with back-plates
that were glued on with an inexpensive non-conductive adhesive. I think
the castings had a lip that surrounded the outer edge of the back-plates
that was flared-in a bit with some sort of mechanical press to improve
electrical contact before the glue was applied. The higher quality
passives were tin-plated so that back-plates could be solder-sealed.
These solder-sealed passives radiated significantly less than the ones
that had the glued-on backs. Likewise system operators were always
looking for connectors that were easy too install and provided
consistently good connections with low "ingress/egress". Nevertheless,
the radiation from a run-of-the-mile F-connector might be 90 dB down
from an isotropic radiator. A better connector might be down 115dB (I
don't remember the exact numbers, but that was the order of the
magnitudes involved)
So yeah, connectors can radiate, but not very much. If you don't believe
me, hook a good shielded 50 ohm load to the RX input of your HF receiver
and measure the noise floor. Then hook up a long run of coax comprised
of multiple sections of coax spliced together with multiple barrel
junctions to your HF receiver input and put the 50 ohm load at the
far-end of that run. In most cases if you are using properly-installed
decent-quality connectors and adapters, you won't hear much, if any,
change in the noise floor. And of course, the minute you loosen the
connector on the back of the receiver and pull it out a little so only
the center conductor is making contact, the noise floor will (depending
on where you live) go up something like 20 or 30dB as that long run of
coax is converted from a shielded enclosure into an antenna.
73, Mike W4EF......................
On 10/1/2025 3:09 PM, Jack Brindle via TowerTalk wrote:
Jim, I’m going to give you the same answer you gave me. "I don't buy any of
this.”
In a perfect world, where connectors perfectly match the characteristics and
construction of coax, I would agree. We don’t live in that world.
The proof? The impedance bump reported by a TDR when it “sees” a connector. Our
sensors are telling us that something is up there that doesn’t agree with your stance. In
fact we have a lot of evidence that something else is going on from that and other sensors.
Like Wes, I would love to see a good mathematical analysis that shows what happens at
a connector, and what happens at the very end of the coax shield. That should be quite
revealing. I don’t remember such an analysis from my Fields and Waves class
from oh so long ago. I do remember the treatment of theoretical coax, and being asked
about it on tests. Alas, there was never any discussion of coax activity at
terminations or on connectors, theoretical or not.
73,
Jack, W6FB
On Oct 1, 2025, at 3:34 PM, Jim Brown <jim@audiosystemsgroup.com> wrote:
Very well put, Joe. Exactly right.
There's another issue at play too-- shielding effectiveness based on the
quality of the shield. It's quantified as the Transfer Impedance of the shield,
defined as the ratio of the differential voltage induced by shield current
divided by that current. The lower that number, the better the shield. The
lower limit is the resistance of the shield at the frequency of interest.
Factors that affect it are the shield construction, like the weave of braid,
the combination of foil and braid. One of the major virtues of hard line is
that the shield is solid. That's also why cables are made with dense double
braid shields silver coated copper.
Years ago, shielding effectiveness came up in work we were doing in the EMC WG of
the AES Standards Committee, and I found a book by Anatoly Tsaliovich on the topic,
who was at AT&T Bell Labs when he wrote it.
73, Jim K9YC
10/1/2025 12:43 PM, Joe Subich, W4TV wrote:
On 2025-10-01 2:51 PM, Wes Stewart via TowerTalk wrote:
At the very end of the cable (or connector) there is no inside and outside of
the outer conductor, there is just the conductor, hence
there is no skin effect at that point.
This is only true if the shield is simply "cut" as in the case of
the coaxial vertical. If the cable is terminated in a connector
- either soldered or crimped - the finite thickness of both the
shield and the connector will maintain the two wire behavior of the
shield through the "splice" so long as the shield and connector
are more than 'n' skin depth thick at the operating frequency.
Even in the case of a braided shield, RF flows *on the surface* -
it does not "weave back and forth" with the braid. This is one
reason that "hardline" and cables with a second foil shield have
lower losses than equivalent size size "double braided" cables.
Common mode currents - unbalanced currents on the exterior of
the shield - are an electromagnetic phenomena and only possible
because RF fields force the current to the *surface* of the
shield - either the outer surface for externally applied (common
mode) fields or the inner surface for differential (transmission
line mode) fields.
The only time those currents are combined is when the transmission
line is interrupted - e.g. the shield is formed into a pigtail -
at an antenna or when brought into equipment without proper
concern (design) for "pin 1" issues.
In any case, common mode currents can be present in non-coaxial
lines. Even simple "zip" cord or other parallel lines can be
treated by applying an impedance to the unbalanced circuit (as
is quite common in noise suppression applications).
73,
... Joe, W4TV
On 2025-10-01 2:51 PM, Wes Stewart via TowerTalk wrote:
Jim,
I think you're missing Jack's very interesting point. I've used an open ended
cable as an example, but a mated pair of your favorite connectors is no
different.
At the very end of the cable (or connector) there is no inside and outside of the outer
conductor, there is just the conductor, hence there is no skin effect at that point. I'm
not smart enough to figure out how far down the cable the skin effect develops. But this
raises a question in my mind. We've all seen a thousand times the drawing of a coax-fed
dipole, where current is "spilling over" the open end and becoming a
common-mode current on the outside of the cable. A smarter mind than mine needs to
'splain this to me.
Wes N7WS
On Wednesday, October 1, 2025 at 10:28:53 AM MST, Jim Brown
<jim@audiosystemsgroup.com> wrote:
On 10/1/2025 7:46 AM, Jack Brindle via TowerTalk wrote:
Connectors are very important in this system. They must be added to the
analysis. Without them, we have to question the validity of the tests.
No. Common mode and differential mode currents are a characteristic of
transmission lines, and common mode can be present on 2-wire lines if
the system that includes the antenna, the transmission line, and
termination in the shack has imbalance. The mechanism by which common
mode in coaxial line is on the outside of the shield is skin effect, and
it's present in those connectors.
Soldered or crimped, the connector(s) is/are simply part of the
transmission line, carrying the differential and common mode current
that is in that system (antenna, line, shack). Depending on their
construction, they can introduce some discontinuity in the differential
circuit.
73, Jim K9YC
_______________________________________________
_______________________________________________
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
_______________________________________________
_______________________________________________
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
|