Coaxial Cable Leakage


Coaxial Cable Leakage

Home

added coaxial cable measurements July 11, 2012

 

Related pages

Unwanted antenna coupling

coaxial lines and shielded
wires

how a shield works (skin depth)

Station RF grounding

Bench Test Fixture July 2012

Data for common mode current ingress into cables is mostly non-existent. The
data that is available is not very useful, because test method details are not
readily available. A manufacturer might tell us shielding effectiveness is so
many dB, but what does that really mean?

Another problem is the abundance of opinions and advice regarding shields and
shielding, and generalizations about what does what inside a cable, including
functionality or operational contribution of shields and foils inside a cable.
Advice and information in amateur radio circles seems to be based on opinions,
observations, or feelings about VHF or UHF performance, or how audio cables
work.

Thinking about this problem, as it relates to amateur radio, I decided to
build a fixture to directly measure coaxial cable ingress from common mode
shield currents. Building a fixture for lower frequencies is not that difficult,
because groundplanes and shields are very predictable. They are easy to test and
verify.  The concerns for an “outside-in” measurement are:

  • The detector system cannot respond to the outside shield common mode
  • Shield common mode current has to be known
  • Shield common mode impedance from cable end-to-end has to be very low
    (this makes the system stable by confining current to the cable tested)  
  • Center conductor (inside current) has to be known
  • The cable must be terminated by impedances similar to the real-world
    application
  • The system has to be verified to ensure measurements are dominated by
    what we actually intend to measure

 

I constructed the following fixture:

test fixture common mode current shield leakage

 

Loads match cable impedance

RF load current is found with a HP RF level meter across the load

High power RF source is an IC751A through an ATR30 tuner

Current sample is a calibrated clamp-on RF current probe

 

 

I normally set shield current at either 300 mA or 1 ampere. Load current is
measured by voltage drop across a 75-ohm resistor using a high impedance RF
millivolt level meter. This is a significant amount of common mode current.

Verification of Fixture

To verify the fixture, I included a dummy port with load near the measurement
connector. I can verify three ways:

  • With the measurement device on the dummy port, dummy port voltage is
    read with normal cable connections and full cable excitation
  • With the measurement device on the LOAD port and the cable on the dummy
    port, voltage on the LOAD port is read with full excitation
  • With an open shield and normal measurements, which produces a near zero
    dB attenuation    

In the first two cases, my meter stays out-of-lock. Lock occurs with about
0.2 mV, while normal readings (so far) are in the 5-40 mV range. This shows an
insignificant error. In the last case, the fixture shows within 0.5 dB of zero
loss (the error is from calibration or tracking errors). 

 

Measurements

Ratio of outside shield current to center conductor current for 8-ft cable
length:

Cable type 1.8 MHz 3.8 MHz 7.8 MHz
old flooded CommScope foil + braid removed from
underground
-70.14dB -76.16dB -76.94dB
Thick copper RG-6/U unused stored indoor but old -71.48dB -74.48dB -72.24dB
old Dish Net cable non-flooded -66.62dB -68.20dB -69.66dB
New CommScope Bright Wire F6 -80.0 dB -87.96dB -99.1 dB
New CS Bright Wire with braid removed for 7 feet (single foil
only)
-78.26dB not tested not tested
       


Bright Wire link

Removing braid has a very noticeable effect on common mode
impedance, because injected current changes. It does not necessarily change
common mode coupling very much. Even on 160 meters, innermost foil dominates
shielding.

The Dish Network cable was just removed after being outside since 2008. It is
a foil / braid non-flooded cable. It appears to have excessive resistance and
bad connections on the outer braid, because flexing changes RF current about
10%. Flexing does not change ingress very much, the inner bonded foil must be
intact.

Connector Mounting

Connector mounting is critical.  For good CM rejection, cables entering
boxes should mount on a common metallic wall. The wall must have significant
conductor area to maintain low impedance between connector grounds, not
moderately-sized circuit traces or wires. The exception is at an intentional
interface to a shield isolation system.

This is a commonly seen connector wiring method. Techniques like this are
often used where cables enter plastic boxes.

poor connector mounting and wiring

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Here is how a connection method like this, with 3-inches of coax, impacts
measurements of

Cable type 1.8 MHz 3.8 MHz 7.8 MHz
brand new flooded CommScope foil + braid -71.48dB -75.56dB -79.08dB
New CommScope Bright Wire F6 -80.0 dB -87.96dB -99.1 dB
CommScope Bright Wire to connector mounted as shown
above, wired to same test fixture
-64.60dB -58.58dB -52.24dB

Some Conclusions

Interesting conclusions can be drawn from this data, and observing system
behavior.

First, just as it should behave, a closed cylinder or wall, even copper or
aluminum foil, does an excellent job of shielding time-varying magnetic fields. The
magnetic field does not easily pass through the aluminum, as long as the shield
is several skin depths thick.

Second, as long as the frequency is high enough to have at least several
skin depths thickness in the shield or wall, the foil shield nearest the center does all or most of the shielding
work. As long as the foil is intact, common mode isolation stays about the same.

Third, the outer shield affects common mode impedance of the cable outside
much more than it actually affects common mode ingress. When the braid was
removed, excitation power had to be significantly increased to maintain the same
common mode current. This shows a higher common mode impedance of the shield.
Cable ingress did not track this increased impedance. It increased at a much
reduced rate (because of time restraints, I only tested this on 160 meters).

Fourth, connector mounting is critical, especially on higher frequencies,
for best CM (common mode) ingress immunity. Connectors must mount directly
to  low resistance and impedance groundplane paths common to other
connectors, unless the device is intentionally planned for CM shield
isolation. On lower bands, it isn’t so much the total metallic enclosure
sealing, but rather the groundplane impedance between different ground
connection areas.

Real-world Measurements (leakage into a new 3000-foot long dual-shield coaxial
cable) in summer 2006

A
separate receiving
antenna and receiver
can be used to find
new multipliers
while making
contacts on the main
radio. A second radio also can be used for two-transmitter interleaved operation
on one band, provided one transmitter is allowed active at any instant of time.

The purpose of
this test was to see
if cable shield
leakage of the
transmitted signal
would exceed signal
level picked up by
Beverages or other antennas. The goal of this was to test transmitter signal
ingress into the receiver’s feed line and shack wiring, comparing the level from
unwanted cable and wiring ingress to levels from the same transmitter into
distant receive antennas. This data should also be
useful in other
applications, or just to illustrate the amount of ingress through this type of
common, inexpensive, coaxial cable.

This
line is CommScope
dual-shield F11
(roughly RG-11 size) cable.
It consists of a
single 100% foil
shield with a single
60% coverage
aluminum braid
overlay. This is a flooded cable for direct burial, with snap-and-seal
connectors.

 

Building
Entrance (old pictures)

 

Building
Entrance Panel

 

 

Test Conditions

200 watts
transmitter power.
Cable terminated in
75 ohms. Signal levels are measured
at receiver.
Antennas used were
my most distant
beverage antenna group approximately 2500 feet (one half mile) from my
transmitting antennas. “dBm” values below for RX
antenna dBm are the values with
actual transmission
line losses
included.

160 Meters

Transmitting
Antenna
Strongest
RX antenna signal level of transmitter in dBm
Weakest case RX
antenna signal level of transmitter in dBm
Terminated
RX feed line  level dBm
1/2 wave dipole, 130
ft high, immediately
above and parallel to receiving antenna coax
-25 -41 -38
High Dipole 300
feet high above
coax
-23 -39 -57
Vertical 200ft
high 350 ft from
coax
-14.5 -34.5 -44

80 Meters

Antenna Strongest
RX antenna signal
from TX in dBm
Weakest TX signal from
RX
antenna in dBm
Terminated
RX feed line level dBm
Low 80M dipole parallel and 40 ft
above receiving antenna
coax
-43 -52 -45.5
High 80M dipole 155
feet above
coax (parallel to RX feed line)
-34.5 -55 -56
High 80M dipole 150
feet above coax
(right angle to RX feed line)
-26.5 -36.5 -44

There are only
two cases where
signal ingress can
exceed signal level
from the beverages
that are loacted one half mile away.
Both cases are where
a low dipole is
mounted right above
and parallel to the feed line. In
these cases, the
feed lines from the
low dipole test antennas also
paralleled the F11
receiving antenna feed line
for at least 200
feet in the same bundle.

160 Meters

Antenna Worse-case leakage headroom
Low dipole
130 ft above receiver
feed line
-3dB
High Dipole
300 ft above receiver
feed line
+18dB
Vertical
Omni
+9.5dB

80 Meters

Antenna Worse case leakage headroom
Low dipole
40 ft above
receiver
feed line and
parallel with
feed line
-6.5dB
High
Dipole155 ft 
above receiver
feed line (and
right angle to
Beverages)
+1dB
High dipole
150 ft above
receiver
feed line and
right angles to
feed line
(broadside to
Beverages)
+7.5dB

 

Conclusion

Even with
Beverages 1/2 mile
from the
transmitter, there
are very few cases
where signal ingress
into the coax
exceeds the signal
level from the
Beverages. The only
cases of failure
were when low
dipoles parallel to
the feed line were
excited and compared
to a Beverage with a
deep null towards
the transmitting
antenna. This data
assumes good coaxial
connections. Common
mode chokes were not
used in this test.

Certainly for any
beverage
installation closer
than 1/2 mile to
transmitting
antennas feed line
ingress will not be
a problem. All of the signal will be from the receiving antennas and not leakage
through the coax shield. 

Hit Counter
since Jan 12, 2007