Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems

565

ESTIMATION OF THE IMPACT OF WIDESPREAD INSTALLATION OF ANTENNA SELECTOR

SWITCHES ON SIGNAL LEAKAGE FROM CABLE TELEVISION SYSTEMS
David J. Large
Gillcable TV
San Jose, California

INTRODUCTION Should the customer education effort be

successful, the result will be that
As part of its regulations governing many cable-served television sets will be
obligations of cable television, the switchable between cable and antenna
Federal Communications Commission has signals, thereby assuring maximum
traditionally required carriage of local programming choice for consumers. The
television broadcast stations. necessary byproduct is that, in those
Recently, however, those rules were cable homes, an efficient radiating and
overturned on constitutional grounds.1 receiving element will be brought into
Subsequently, the Commission has drafted close electrical and physical proximity
new rules with less stringent carriage to the cable. Furthermore, a significant
requirements for an interim period of percentage of these switches may be
five years and no requirements after installed by technically unsophisticated
that time.2 cable customers.
In order to protect the viewing choices The purpose of this study is:
of television watchers, the rules as
released in November 1986 required that, (1) To determine the effect of such
for new installations, switches must be installations on cable operator's
installed at every TV set to allow obligations to control leakage of
selection between cable and antenna their signals into the atmosphere
inputs and that cable systems must try and, coincidently, the potential
to pursuade existing customers to either impact on other radio services.
take a free switch for do-it-yourself
installation or pay the cable system's (2) To determine the degree to which such
labor cost for installing one. installations may adversely effect
the quality of reception afforded
Acting in response to numerous Petitions customers because of antenna-received
for Reconsideration, the Commission in local television signals causing co-
March 1987 revised the selector switch channel interference with cable
requirements so that installation of signals.
switches would be at the option of the
customer, who would be required to cover The various technical and practical
the cost involved. Under these revised results of limited cable-antenna
rules, customers would also be permitted isolation will be examined and, in each
to purchase switches from third party case, a minimum required isolation will
vendors and install them. The mandatory be calculated. Then, several typical
requirements for customer education were installations using a variety of hardware
maintained. Finally, acting in response and configurations will be measured to
to the material presented in this determine actual isolation achievable in
document (among others), the Commission practice. This data, combined with
announced an intention to issue a statistical data on the relative
separate Notice of Proposed Rule Making quantities of various types of hardware
to study the signal leakage issue. in use will be used to predict the

0098 3063/87/1100 0565$01.00 © 1987 IEEE

Manuscript received May 11, 1987.
566
relative difficulty of meeting the
derived minimum required isolation.
Since the overall effect of some of the
leakage mechanisms is dependent on local
conditions, the community of San Jose,
California will be used as a case study.
The formulas presented should allow
other situations to be predicted,
however.
REQUIRED ISOLATION
DISCUSSION OF ISSUES:
Limited antenna-cable isolation results
in both egress of cable signals and
ingress of off-air signals. The
practical results of excessive egress
are:
(1) Interference (caused primarily by
individual leaks) to other radio
services using portions of the
spectrum other than the VHF
broadcast television channels.
(2) Co-channel a n d / o r ' g h o s t'
interference to non-subscribers,
again primarily caused by individual
leaks.
(3) Interference to aircraft navigation
and communications systems caused by
the aggregate effects of many small
leaks in the airspace above cable
systems.
Excessive ingress results in co-channel
and/or 'ghosting' interference to
subscribers on VHF channels 2-13 and
various forms of picture impairment from
other radio services on mid- and super-
band channels.
FCC rules spell out limits on both
individual leaks and the cumulative
effects of total system leakage. Part
15.606(a) specifies the required
isolation of antenna selector switches,
Part 76.605(a)(11) specifies the maximum
allowable individual leak magnitude,
Part 76.611 specifies the maximum total
system leakage (CLI) and Part 76.613
contains the general definition
forbidding any leakage that "..seriously
IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
degrades, obstructs or repeatedly
interrupts . . ."1 other communications.
ISOLATION REQUIREMENTS OF PART 15:
The rules specify that the switch must
have sufficient isolation that the signal
fed back into the antenna used is
< 0.346 aR microvolts where R is the
impedance of the antenna lead. In a 75
ohm system, the maximum voltage is 2.9964
microvolts (twice that in a 300 ohm
system) and the resultant power leel is
1.1972 X 10- 10mW = -50.468 dBmV.3'
It is important to note that the language
of Part 15 does not apply to the switch
as an isolated element, but to the switch
as part of the total system, since the
required switch isolation is obviously a
function of input signal level.
Assuming that the level of cable input
signals can vary from 0 to +15 dBmV, the
cable system will have to be isolated
from the antenna by 65 dB to meet this
criterion.
The author feels that the adequacy of
these rules as applied to cable-antenna
switching should be examined. The
situation is significantly different from
that which was contemplated when the Part
15 rules were drawn up. In a typical
application for an 'antenna selector'
switch, the device used (whether VCR,
computer or video game) is a single
channel device operating on a low-band
VHF channel that is not used in the local
area for off-air reception. Both ingress
and egress interference is minimal as
there is no common channel. Cumulative
effects are minimized by:
(1) The low population density of such
devices
(2) Their occasional, rather than
continuous, use
(3) The energy dispersal caused by loose
frequency tolerances on modulators
within home equipment
In the cable situation, almost the entire
VHF spectrum is used simultaneously and
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems 567

continuously including off-air occupied Under the latest rule revision, the
frequencies. Furthermore, the density Commission has apparently extended the
of possible efficient radiating elements definition of 'system component' to
in a community is very high, being the include subscriber-owned equipment (the
product of television set density and final language had not been released at
cable penetration. Finally, all the the time of the drafting of this
radiating elements are fed from a single article).
coherent source. Thus, both individual
and cumulative effects of small leakages
is likely to be more severe. If the radiator has dipole efficiency, we
we may use equation [2] to calculate the
power necessary to reach that field
ISOLATION REQUIREMENTS OF PART 76.605: strength:
The measured field strength from a
transmitting dipole antenna may be 2
calculated as: p O225gv/m 1 1
= 137.6 52.8IJ
7.58 X 10- 8 mW = -22.5 dBmV
E = 137.6 / [1]

In the more typical case of a rooftop

antenna with 10 dB gain, the input power
where: E is the measured f ield would have to be limited to -32.5 dBmV to
strength in mV/m at 1 mile avoid a violation. Given a maximum cable
signal level of +15 dBmV, the isolation
P is the transmitted power required is 47.5 dB.
in KW5

ISOLATION REQUIREMENTS TO MEET CLI:

Since field strength varies inversly
with distance, this may be scaled and Of equal interest is the effect of many
restated as: small leaks on the cumulative leakage
(CLI) of entire cable systems. We may
approach this problem theoretically as
follows:
E = (137.6/D) 4F [2]
(1) Calculate the average transmitted
power density (due to limited
where: E is the measured f i el d isolation) per square mile of the
strength in pV/m system.
D is the distance in miles
(2) Calculate, for an airplane passing
P is the transmitted power over the center of the system at a
in mW height of 450 meters, the measured
field strength due to the radiated
power from an incremental portion of
Part 76.605(a)(11) of the Commission's the system.
rules require that the field strength of
individual leaks of cable signals (in
the range of 54-216 MHz) be below 20 (3) Integrate this effect over the entire
,u V/m measured at a distance of 10 feet system, adding og a power rather than
from the 'system component'. a voltage basis.
568 IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987

Diagramatically: need to add the power effect of the

individual voltages, i.e.:
Er

or, in the limiting case:

\D Et tEtti
I 0i:1 R(137.6)2
D2 Pr

y27rR (137 rdrdO

oJ (H2+r2,)rdd
P
To
1r27Tr R
137.6 dJJ 2 drdO

Where: R is the system radius in R2

243.9 Pd n 1+
miles [3]
H is the airplane height in
miles

D is the distance from an Where Ettl is the total resultant

incremental source to the measured field strength in p,V/m due to a
measurement point uniformly distributed field of dipole
radiators fed with an average power
= VH2 + r2 density of Pd mW/mi2 if the measuring
point is located H miles above the center
Pr is the incremental power of a system of radius R miles.
radiated = Pd r a re
For the San Jose system, the average
where: Pd is the average number of homes per square mile,
transmitted power including non-residential areas, is about
density in mW/mi2 3014. With a system penetration ratio of
50%, thi.s gives 1507 serviced residences
r is the distance from per square mile. Approximately 20% of
the center of the these residences have more than one set
system to the on service, so the number of potential
incremental source radiating sources per square mile is 1.2
in miles. X 1507 = 1808. Thus, the average power
density will be 1808 X Pc where Pc is the
Then, from equation [2] above, Er = average leakage level from an individual
(137.6/D)Iv where Er is the measured set.
field strength, measured in pV/m, due
to the incremental source Pr. The system radius is approximately 5
miles = R, so that the total number of
In order to integrate the effect with potential radiating sources is 1808 X ir X
power rather than voltage addition, we (5)2 t 142,000.
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems
The height of the airspace measurement
for CLI is 450 meters above average
terrain level so H = 0.280 miles.
For this H value, equation
to:
Etti = 243.9 d 1+]
Customer antenna leakage represents a
heretofor minor reradiation mechanism.
In order that this new source of leakage
not add significantly to the total
system leakage, it should not exceed 10%
of the total allowable field measurement
(10 pV/m) or 1 ,uV/m.
Solving equation [4] for Pd we get:
Pd= [3n('+ 007)]
[243]
Using the numbers for the San Jose
system, we can calculate the maximum
allowable power density:
P= 2.912 X 1o-6 mW/mi2
Given a radiator density of 1808/mi2,
the average power delivered to each
radiator cannot exceed 1.61 X 10 9mW =
-39.2 dBmV. If the average drop level
is +5 dBmV, then the average isolation
required is 44.2 dB between cable
signals and antenna lead for universal
switch instal lation.
EFFECT OF R&NDOM WIRING ERRORS ON CLI:
Given that customers will have the
option of installing their own switches
or paying the cable operator to install
them, we need to consider the
probability of wiring errors. For
example, of the six possible ways to
install a high-isolation A/B switch
(having three female type-F coaxial
569
connectors), four result in a direct
cable-antenna connection in one switch
position. Additionally, some homeowners
may 'simplify' the installation by
[31 reduces connecting the antenna leads directly
across the TV set terminals in parallel
with the cable input. While a good
subscriber education program should
minimize the number of such errors, it is
worthwhile to know what error rate is
possible without seriously affecting CLI.
We can calculate the percentage of such
errors that would have to happen to
create a serious CLI situation as
follows:
From [6] above, the maximup allowable
power density is 2.912 X 10 mW/mi . The
radiated power for a single leak of +5
dBmV level is 4.22 X 10-5mW (resulting in
a 472 PV/m field strength ground-based
measurement at 10 ft distance), so that
the average number of leaks must be below
(2.912 X 10-6)/(4.22 X 10-5) =
0.069 leaks/mi2
[5]
Given the radiator density of 1808/mi2,
the maximum percentage wiring error is
(0.069)/(1808) = 0.0038%
or 5.5 out of San Jose's potential
142,000 radiating sources.7
[6]
If just 1/10 of San Jose's subscribers
opted to install switches, the percentage
of allowable errors would only increase
tenfold to 0.038%.
It should be noted that at the maximum
drop level of +15 dBmV, the ground-based
measurement at 10 ft would yield a field
strength of 1492 ,uV/m from a dipole
radiator or 4718 pV/m from a 10 dB gain
rooftop antenna-sufficient to cause
significant disruptions to other users of
the spectrum and serious non-subscriber
co-channel interference over a wide area.
For a discussion of these predictions in
relation to the conclusions contained in
the Final Report of the Advisory
Committee on Signal Leakage, see
Appendix .
570 IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
ISOLATION REQUIRED TO AVOID VISIBLE
PICTURE IMPAIRMENTS:
Allowable Desired/Undesired Signal
Ratio:
Key to determining the degree of
isolation necessary to assure that the
visible effects of ingress and egress
are acceptable is the determination of
the minimum desired/undesired carrier
level.
Part 76.605(a)(8) of the rules for off-
air co-channe-l specified a carrier ratio
of 36 dB. That limit, however was
predicated on the precise 10 kHz and 20
kHz offsets that are used in broadcast
frequency allocations for the purpose of
limiting the visible effects of
interference. Furthermore, the rules
specify a maximum tolerable situation as
opposed to a situation of no noticeable
degradation. Finally, the threshold of
visibility has decreased in recent years
with improvements in television sets and
larger screen sizes. In comparison to
the off-air situation, cable frequencies
may vary from nominal assignments due to
frequency tolerance and stability by as
much as +/- 25 kHz as specified in Part
76.605(a) (2).
The visible effects of ingress co-
channel interference may be reduced in
some cases by phase-locking the
processors and modulators in a cable
system to the local off-air stations.
Many operators do this to avoid
complaints due to direct pickup of
strong station signals in the internal
wiring of TV sets. It should be pointed
out, however, that systems using HRC and
IRC channelization schemes do not have
that freedom as all of their channels
are lockgd to a single frequency
reference.
The author has directed tests to
determine the threshold of visibility of
an interfering carrier in the past. If
the frequency of the interfering carrier
is allowed to vary to the point of
greatest interference, it will be
visible down to a level of 60-65 dB
below the desired video programming.
Tests were also conducted in which the
actual off-air signals were combined in
controlled levels with cable programming
to determine the threshold of visibility
under various conditions with the
following results, expressed as the
maximum carrier ratio in dB for which co-
channel beats were clearly visible:
A. Same video, carriers phase locked
together: 34-41 dB
B. Same video, carriers not locked
together: 58- >60 dB
C. Different video, carriers phase
locked together: 36-48 dB
D. Different video, carriers not locked
together: all >60 dB
Which confirms a minimum necessary
carrier ratio of 60 dB in non-locked
situations and 40-45 dB with phase-locked
RF carriers.
Isolation Necessary to Avoid Ingress Co-
channel Interference:
In an area with strong local TV signals,
the antenna downlead siynal strength can
easily be +20 dBmV. If the cable
signals can vary down to 0 dBmV, then a
switch isolation of 80 dB would be
required to keep antenna-conducted co-
channel signals 60 dB below cable
signals.
As a practical check on the isolation
necessary in practice, the antenna signal
level from a typical rooftop antenna was
measured in San Jose, along with the
attenuation necessary to eliminate
visible interference when combined with
0 dBmV cable signals (all carriers were
phase-locked to respective off-air
carriers) (see TABLE 1).
Thus in San Jose, an attenuation of
nearly 60 dB is necessary to protect
against the San Francisco stations over
40 miles away, even with carriers phase-
locked.
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems 571

This data also confirms our field subscriber's antenna would have to be no
experience with so-called video greater than 0.225 ,uV/m for channels 2-6
switching networks which perform the and 0.56 iiV/m for channels 7-13, based on
necessary RF switching with a converter a required 60 dB level difference.
and VCR. In that case, the RF signal
levels are similar, but carriers are not The most serious potential non-subscriber
locked. We have found that isolations interference situation results when the
of 70-80 dB are essential to avoid radiating source is between the non-
subscriber complaints with those subscriber's antenna and the station.
devices. For single family residences I have
assumed an antenna-to-radiating source
Isolation Necessary to Avoid Interfering spacing of 100 ft.10 In San Jose, local
with Non-Subscribers: stations are located to the North, East
and South of the system so it is quite
To the extent that cable signals possible for the rooftop antenna of the
reradiate, they not only cause technical interfering source to be aimed directly
violations of the Commission's rules, into the major lobe of the non-customer's
but may interfere with other radio antenna.11 Given the limited directivity
services and non-subscriber's television of most home antennas, it is certainly
reception. not unreasonable to assume at least
dipole gain in the direction of the non-
The Commission has defined the Grade-B subscriber's antenna.
limit field strength of a TV station as
225 pV/m for channels 2-6 and 560 ,uV/m In that case, we can use equation [2] to
for channel 7-13. To avoid visible co- solve for the maximum allowable radiated
channel interference, therefore, the power that will not cause visible co-
field strength of the reradiated cable channel interference with worst-case
signals as received at the non- frequency offsets as follows:

TABLE 1

OFF-AIR CHANNEL LEVEL CABLE CHANNEL ATTENUATION NECESSARY

2, San Francisco +12 2A-same video 46 dB

2, San Francisco +12 2B-different 56 dB
4L, San Francisco +5 4A-same video 40 dB
4L, San Francisco +5 4B-different 44 dB
5,v San Francisco +10.5 5A-same video 48 dB
5,2 San Francisco +10.5 5B-different 55 dB
7, San F rancisco +17 7A-same video 49 dB
7, San Francisco +17 7B-different 57 dB
8, San Jose +25.5 8A-different 80 dB *
8, San Jose +25.5 8B-same video 80 dB *
9, San Francisco +15 9A-different 48 dB
9, San Francisco +15 9B-same video 46 dB

(* pickup directly in monitor makes results questionable)

572
For channel 2: [20gV/m( 1 2 (6) Isolation necessary to avoid causing
=
137.6 '528)j
= 9.59 X 1010
-41.4 dBmV
ro.5 6pv/m/ 1
For channel 13: P = [13.6 52.8
= 5.94 X 10-9 mW =
-33.5 dBmV
Thus, to avoid such interference to a
low-band station with a maximum drop
level of +15 dBmV, an isolation of 56.4
dB will be required. With an average
drop level of +5 dBmV, the required
isolation is 46.4 dB.
SUMMARY OF ISOLATION REQUIREMENTS:
In summary, then, if the result of a
large percentage of cable customers in
San Jose having rooftop or 'rabbit ear'
type antennas connected via switches to
their TV sets is not to cause a
significant increase in leakage, the
following numbers must be maintained:
(1) Isolation to meet Part 15.606 at
maximum drop level:
average drop level:
(2) Isolation to meet Part 76.605 at
maximum drop level: 47.5 dB
average drop level: 37.5 dB
(3) Average isolation for 10% CLI
contribution with
100% of sets wired: 141.2 dB
10% of sets wired: 314.2 dB
(4) Percentage 'wiring errors' for 10%
CLI contribution with
100% of sets wired:
10% of sets wired:
(5) Isolation necessary to avoid ingress
co-channel interference from local
VHF's at
minimum drop level: 80 dB
average drop level: 75 dB
IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
co-channel interference to Grade-B
low-band channel at non-subscriber at
mW = maximum drop level: 56.4 dB
average drop level: 46.5 dB
ACHIEVABLE CABLE TO ANTENNA ISOLATION
2 LABORATORY TESTS:
In order to evaluate the difficulty of
meeting these isolation criteria, the
Gilloable (San Jose) Research and
Developement lab conducted tests on
connection permutations typical of
various home TV situations. The test
equipment used was a Wavetek 1801A
sweep/signal generator and a Tektronix
7L13 spectrum analyzer. Except as noted,
all cables were terminated in their
proper impedance and cables were widely
separated to minimize incidental
coupling.
Basic Switch Isolation Tests:
For these tests, a signal was applied to
one input of the switch under test, the
armature was terminated and the signal
level was measured at the other input
port. Where impedance matching was
required for 300 ohm terminals, a CATV
Services model TK-55 matching transformer
was used.
65 dB
55 dB Tested switches were:
(1) Pfantone TV-SW300 Cable/Antenna
Selector Switch with 300 ohm antenna
and TV terminals and 75 ohm coaxial
cable connector.
(2) EIE/RCA AB7-75 A/B switch with three
F-female connectors
(3) Arvin Systems Model 600B A/B switch
with three F-female connectors
0.0038% The results were as follows:
0.038%
Isolation
l100-150MHz 200-400MHz
(1) Pfantone
TV-SW300 50 dB 42 dB
(2) EIE/RCA A/B7-75 >90 dB >90 dB
(3) Arvin Model 600B >90 dB 87 dB
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems 573

It thus appears that either of the high (2) A Sony Model CVM1250 which differs
quality coaxial switches are capable of from the above in that it has both 75
achieving the isolation required, while ohm coaxial and 300 ohm screw
the Pfantone switch measured terminals available. Internally, a
significantly less. twinlead jumper is still used.
Situational Testing: (3) A Sony Model KV-1976R which has only
coaxial input terminals and is
In a practical installation, the total furnished with a plug-on balun for
isolation between the cable input and 300 ohm feedlines.
customer's antenna is a function of the
switch used, external components In all cases, the input signal was
(matching transformers, connectors, generated by the Wavetek generator and
cables, etc.) and the TV set itself. To connected via coaxial cable to the
assess the contribution of these switch. The antenna lead was a 10 foot
external components, Gill examined the section of 300 ohm twinlead, routed away
isolation of various actual installation from the set and terminated with a
configurations. matching transformer into the Tektronix
analyzer. If necessary, a matching
Three TV sets were tested: transformer was used at the switch (as
would be used in a normal installation)
(1) A GE model WMT153SAVE with 300 ohm (See TABLE 2).
screw antenna terminals,
representative of many older sets in Summary of Test Data:
our system in that it has an
internal 300 ohm twinlead jumper From the above it appears that it is
from the non-metallic back cover to unrealistic to expect customers' antennas
the tuner. to be isolated from cable signals (and

TABLE 2
Isolation
100-150MHz 200-400MHz
TV#1 ,Switch #1, leads carefully separated 47 dB 34 dB
It .I leads parallel to floor (3') 49 dB 24 dB
it it leads loosely twisted for 3' 40 dB 26 dB
TV#1 ,Switch #3, leads carefully separated 54 dB 40 dB
it it leads parallel to floor (3') 40 dB 36 dB
tt it leads loosely twisted for 3' 43 dB 34 dB
TV# 1,Switch #2, leads carefully separated 60 dB 50 dB
1t I
leads parallel to floor (3') 47 dB 28 dB
it it
leads loosely twisted for 3' 42 dB 32 dB
TV#1,Switch #2, as above with 1 loose balun lead 28 dB 28 dB

TV#2,Switch #3, using coaxial input, leads parallel to floor 30 dB 36 dB

"t It
As above but with external balun 22 dB 12 dB

TV#3,Switch #3, using coaxial input, leads parallel to floor >75 dB 70 dB

It It As above but using external balun and
plug-on transformer 50 dB 45 dB 12
574 IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
vice-versa) by more than 30-35 dB
average, except in the case of all-coax
sets. In the latter case, 45-50 dB may
be achieved with baluns and 70 dB with
direct connections. The use of high
quality switches of 90 dB isolation
helps, but does not solve the problem in
most configurations because of the
coupling of external components.
A preliminary survey of a sample of San
Jose subscribers sets has shown that 19%
have only 300 ohm terminals, 30% have
both and 51% have only coaxial input
terminals. Thus, approximately half of
the sets in use in that cable system
apparently lack the shielding necessary
to allow adequate cable-antenna
isolation. The bulk of the remaining"
sets are capable of achieving reasonable
isolation if matching transformers are
removed.
Discussion of the Causes of Low
Isolation:
First, most older sets were designed for
connection to a 300 ohm feed line from an
external antenna. Since shielding
against reradiation of antenna-conducted
signals was of no concern, the back of
the sets were non-metallic and an
additional piece of twin lead was used
to connect the screw terminals to the
tuner. This lead is typically fairly
long to allow for removal of the set
rear cover.
Second, most TV tuners exhibit an
approximate impedance match to the
antenna feedline at only the station
currently tuned. At al l other
frequencies the lead is severely
mismatched. The result is that a high
VSWR exists on the feedline and it
becomes a radiating element.
Third, the external antenna feedline
would exhibit reasonably low radiation
characteristics if properly matched and
if operated under ideal conditions,
however it:
(1) Is imperfectly terminated at both
antenna and matching transformer.
(2) Is not straight
(3) Is not spaced away from other objects
(4) Is very close to other radiating
elements near the TV set
Fourth, the matching transformers are
designed for low cost mass production and
are neither truly balanced nor well
shielded.
Finally, minor errors such as loose or
broken balun leads cause only minor
picture impairments (particularly if drop
signal levels are high), yet cause
serious mismatches and consequently high
radiation.
The diagram below illustrates some of the
major mechanisms responsible for low
cable-antenna signal isolation.
Internal
Coupling
Paths 300Q Jumper
300o Antenna
Feed Line
BALUN
Cable Input
IMPLICATIONS OF TEST RESULTS WITH RESPECT
TO REQUIRED ISOLATION
Gill's test results have shown that in
roughly half of San Jose cable
installations, the addition of an antenna
selector switch and connection of an
antenna will result in only 30-35 dB of
isolation. In the majority of the
remaining cases, we would expect 45-50 dB
of isolation and, in a few cases of all-
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems
coaxial installations without matching
transformers, we may achieve 70 dB.
In some cable systems, the low isolation
will not be such a major factor because
some or all installations require the
use of a converter or descrambler ahead
of the television set so that the full
cable spectrum is not delivered to the
antenna terminals. In the San Jose
case, however, about 2/3 of the
customers opt for Basic service only and
directly connect to television sets and
V CR's. Even in the case of Premium
customers, Gill, along with many others
in the industry, provides a switching
device to allow direct broadband
connection of the cable to the
tel ev ision set in an effort to be
'friendly' to the use of extended tuning
range sets, flexible VCR usage, etc.
The result is that, at any one time, the
vast majority of television sets in the
San Jose system are directly connected
to the cable input despite the fact that
Gil l uses advanced add ressab le
technology.
The implications of the achievable
isolation are:
(1) That the requirements of Part 15.606
will often not be met except in
direct coaxial connection
situations.
(2) That average leakage levels of
approximately 20 pVV/m (the limit of
Part 76.605(a)(11)) will routinely
occur with +5 dBmV drop levels for
about half of the sets in use.
(3) That if selector switches are
routinely installed in San Jose,
the CLI contribution from customer
antenna radiation alone will result
in aircraft monitored levels of
approximately 2 pV/m, independent of
any wiring errors, loose fittings,
failed switches or other abnormal
conditions. 13
(4) That a very small percentage of
wiring errors that result in direct
antenna-cable connections will cause
a significant increase
575
system radiation, leading to
potential aircraft radio
interference. The gradual infusion
of better shielded television sets
into the system will not improve this
situation. In addition to their
effect on CLI, such errors and
failures will cause major violations
of Part 76 leakage specifications,
widespread co-channel interference to
non-subscribers and interference to
other radio services. In fact,
Gill's experience with the current
occasional incidence of such errors
is that interference is sometimes
found up to several blocks away from
the offending antenna.
(5) That visible ingress co-channel
interference will occur with antenna-
received signals as low as -25 dBmV
in TV sets using unshielded internal
jumper cables. In other words,
virtually all signals of usable
signal strength will cause picture
impairment to subscribers trying to
view cable signals. Even if the
cable system uses phase locking to
synchronize cable carriers to
equivalent off-air carriers, signals
as low as -5 dBmV will cause visible
impairments.
(6) That even average drop level
installations may cause co-channel
interference to nearby non-
subscribers viewing stations near
their Grade-B signal limits unless
the subscriber has a highly shielded
coaxial installation. At maximum
drop levels, such interference is
likely.
SUMMARY AND TECHNICAL PROPOSAL
An analysis of the isolation required to
maintain high quality subscriber
reception and acceptable levels of signal
leakage shows that it is inconsistent
with the practically achievable results.
A/B switch installation will result in a
reduced level of service to subscribers
and non-subscribers alike and the
like 1 ihood of an unmanageab l e
in total interference level to other radio
576 IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
services, including those with direct
ties to public safety.
The Commission's intent in promoting
freedom of choice and diversity of
programming for the viewing public is
laudable. The wholesale installation of
antenna switching devices, however, is
inconsistent with Commission efforts to
control leakage of cable signals.
We have doubts that it is possible to
keep the incidence of wiring errors and
equipment failures to an acceptable
level. If switches must be installed,
however, we would suggest the following:
(1) That.all equipment that may be
attached to a cable system be
required to meet the signal leakage
requirements of Part 76.605(a)(11).
The Commission has before it now the
Further Notice of Proposed
Rulemaking in Gen. Docket No. 85-301
which specifically addresses that
subject.
(2) That cable operators be required to
offer to install antenna selector
switches at cost on all new
installations if the equipment meets
the criteria above and if the
customer's antenna is equipped with
a 75 ohm coaxial downlead of a
quality at least equivalent to one
of the medium CATV grades.
(3) That all switches used for antenna
selection, whether built into
consumer equipment or external
accessories, be required to meet an
isolation specification of 80 dB
from 54-550 MHz. Such switches
should be provided with coaxial
connections only (to discourage
installations in which external
coupling limits the effective switch
isolation.)
(4) In recognition of their lower
radiating efficiencies and shorter
feedlines, that 'rabbit ear'
antennas be allowed to be connected
to switches as an alternative to
external antennas with shielded
transmission lines.
(5) That cable operators inform existing
subscribers of the availability and
reasons for using selector switches
and install them on request and for
reasonable fees provided that the
customer's equipment and antenna meet
the same requirements as for a new
installation.
This will assure that the Commission's
goals for diversity are met without
causing degraded reception for any
viewers or undesirable interference
levels, at least at initial installation.
While it is hoped that professional
installation of antenna selector switches
on well-shielded equipment will result in
a manageable leakage situation, we would
recommend that the Commission continue to
review consumer antenna leakage to detect
any dangerous increases resulting from
the new switch installation.
REFERENCES
lUnited States Court of Appeals for
District of Columbia, Quincy Cable TV,
Inc. vs FCC, 1985.
2Amendment of Part 76 of the
Commission's Rules Concerning Carriage of
Television Broadcast Signals by Cable
Television Systems, MM Docket No. 85-349
Released: November 28, 1986.
3The exact meaning of this
specification is not clear in a broad-
spectrum application. I have assumed for
this analysis that the maximum voltage
specification applies to each individual
carrier rather than to the total rms
voltage present at the antenna feedline.
40dBmV = 1 mV in a 75 ohm coaxial
system.
5Frederick Terman, Electronic and
Radio Engineering, 1955
6The rationale for this is discussed
in the Final Report to the Federal
Communications Commission by the Advisory
Committee on Signal Leakage and is based
on the random phasing of received
signals.
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems
7As a check on the above
calculations, we can use equation 121
and assume al l of the radiated power is
fed to one antenna and calculate the
distance required to create a field of
1 ,uV/m:
D = (137.6/E) V wherq P = (4.22 x 10-5)
(5.5) = 2.29 x 10- mW
= 2.08 miles or roughly half the
system radius, a very reasonable
result.
8HRC and IRC are techniques whereby
all channels are phase-locked to a
single source so that second and third
order distortion products fall precisely
on the video carriers of interfered-with
channels.
9In urban areas such as San
Francisco, levels as high as +40 dBmV
are not uncommon.
10This is based on typical suburban
lot sizes. In an urban, multiple
dwelling situation the spacings may be
much c loser.
11Given that the measured signal
from a dipole is V( ,u V)
E(1,uV/m)/.021F(MHz), we can calculate the
total path loss between two subscriber
antennas as: Loss (dB) = 10 log
(1750D2F2) G -G2 where D is the
- this experiment, a small number of
distance in miles, F is the frequency in
MHz and G & G2 are the antenna gains in
dB. Thus the total path loss for two 10
dB gain antennas facing each other at
100 feet distance is only 12.9 dB at
channel 2!
12In the case of Gillcable, it has
been standard practice for several years
to use 300 ohm terminals when available
(even if that requires two baluns) in
order to reduce the shock hazard to
employees and customers in the event of
an electrical failure in the set. The
result is that very few sets in the San
Jose system are installed without
matching transformers.
13Calculated by assuming dipole
efficiency radiation in the vertical
direction for half of the serviced sets.
577
To the extent that rooftop antennas with
significant gain are in use, the vertical
radiation will presumedly be less. On the
other hand, in San Jose, as well as many
other cities, a significant majority of
off-air programming watched originates
from a single point distant from the
city. This results in a network of high-
gain antennas coaxially aimed which might
result in an even higher field strength
in the air over the part of the system
facing the signal sources.
APPENDIX
CORRELATION OF PREDICTED CLI RESULTS WITH
CONCLUSIONS OF THE ADVISORY COMMITTEE ON
SIGNAL LEAKAGE:
On November 1, 1979, the Final Report to
the FCC by the Advisory Committee on
Signal Leakage was released. As this
report has provided a technical basis for
subsequent regulation of cable system
leakage levels and procedures, it is
important to show the relationship of the
data and conclusions contained therein to
the technical predictions in this study.
= A. In Section 4.2(6), of their report
(T echni ca l C onc l us i ons), the
Committee indicates that "At least
under the controlled conditions of
artifically created leaks, each about
40 times greater in power than the
largest leaks discovered in operating
cable systems, was insufficient to
cause harmful interference to ground-
air communications services . .
The supporting data shows that
feeding three approximately resonant
dipoles a level of +36 dBmV caused
the 90th percentile field strength at
450 meters over the system to rise 11
dB above background noise to a level
of -98 dBm (corresponding to an
equivalent field strength of
approximately 8.8 microvolts/meter
according to the antenna gain
calculations in Section 2.2.2.)
Assuming that the signal level fed to the
dipoles were reduced to +16 dBmV, the
expected field strength would reduce to
578 IEEE Transactions on Consumer Electronics, Vol. CE-33, No. 4, NOVEMBER 1987
0.88 microvolts/meter. This differs
from the prediction in this study (5.5
dipoles at +5 dBmV to produce a field of
1 microvolt/meter) by a factor of 9.7
dB, which is within reason considering:
1. the unknown spacial distribution of
the radiators in the FCC tests
2. the inconsistent polarization of the
radiators used in the FCC tests
3. the unknown effects of immediately
surrounding objects on the radiators
used in the FCC tests.
The significant variation between
theoretical and measured field strengths
within the Committee's work (up to 15
dB) is discussed in Section 2.5 and
certainly throws into question the
degree to which those results can be
projected into other situations.
As opposed to the test environment used
in the Advisory Committee report,
radiation from cable customers' antennas
should have relatively strong
polarization consistency due to the
design of such antennas. Furthermore,
in the San Jose system where most off-
air stations come from a single
location, the electrical fields will
tend to be aligned. Finally, cable
customers' antennas are designed
specifically to be efficient radiating
elements and are typically mounted well
away from other metal objects.
Given the above factors, the predictions
in this study are not inconsistent with
the findings in the Report.
B. An important conclusion of the report
is that leaks of magnitude less than
100 microvolts/meter (measured at 10
feet) contribute insignificantly to
overall leakage. This is based on a
statistical study of leaks in a
sample of systems. It is no longer a
valid conclusion if the number of
leaks is increased dramatically as
would be the case with limited
isolation between the cable and a
large number of antennas.
The data presented in Section 3.6 shows a
median system leakage level of 257
microvolts/meter. A single wiring error
could contribute much higher levels than
that, but so could 100 leaks of 26
microvolts/meter amplitude. Given the
number of television sets per mile of
plant, we can no longer expect that the
ratio of minor to major leaks will follow
traditional patterns if antenna leads are
universally brought into close proximity
to cable signals.
C. Other conclusions such as the
probability of interference (3), the
difficulty of operator monitoring (4)
and the difficulty of FCC spot checks
(5) are all based on the same
statistical data. Given the
dramatically different expected
population breakdown of cable leaks
under the new rules, all of these
conclusions should be reevaluated.
Large: Estimation of the Impact of Widespread Installation of Antenna Selector Switches on Signal Leakage from Cable Television Systems 579

BIOGRAPHY
David Large, a fourteen-year veteran of
the cable industry, is Senior Vice
President of Engineering for Gill
Industries, San Jose, California. Prior
to that he had responsibility for CATV
test equipment design for Avantek.
Earlier experience includes microwave
test equipment and telemetry system
design. He holds a BSEE degree from the
California Institute of Technology.
David is a Senior member of both the
Society of Cable Television Engineers
and the Institute of Electrical and
Electronics Engineers and is a member of
the NCTA Engineering Committee. He
heads the Subcommittee on Consumer
Interconnection.
David has published a number of
technical articles and made numerous
presentations, particularly in the areas
of addressability, consumer equipment
compatibility problems and stereo sound.
He holds two patents and has a third
pending.