IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, VOL. PAS-89, NO.

2, FEBRUARY 1970 275

Electrical Interference Aspects of Buried

Electric Power and Telephone Lines
FORREST WOODLAND, JR., SENIOR MEMBER, IEEE

Abstract-This paper discusses the audio frequency noise in- evident that the many joint committees organized and sponsored
duced in telephone lines by nearby electric power circuits, par- by the Edison Electric Institute and the Bell System have
ticularly when the power and telephone companies employ joint performed a problem-solving function of great magnitude and of
buried construction methods. The electrical protection considerations virtual necessity. These committees have, for the past 45 years,
of this type of construction are not included in this paper. Com- turned out a series of some 50 technical reports [31, that have
parisons are made with the usual aerial type of joint construction. become a valuable reference for electrical engineers in both the
The factors controlling the amount of interference produced by
inductive influence are described and evaluated. Recommendations power and telephone fields. However, if the quality of telephone
are suggested for limiting this interference to acceptable levels for service is to continue to improve we must continue to do all we
satisfactory telephone service. can to improve the inductive interference situation.
Telephone companies are experiencing more and more de-
mand for more sophisticated circuits. They must furnish circuits
INTRODUCTION that can transmit data with extremely high reliability. They
THE electrical protection considerations involved in the must furnish circuits that can transmit signals for electric power
"random separation" of jointly buried electric power and companies. Many of these power company signaling circuits
telephone facilities have been thoroughly discussed in [1]. carry extremely important messages related to the functioning
The electrical protection considerations have also been covered in and control of the electric power systems. At the same time that
detail in [2]. This paper will be concerned almost exclusively they see a need to improve the quality of transmission on these
with the factors and considerations that affect the amount of circuits, they see that the adverse effects of electric power
noise induced in telephone lines due to the proximity of multi- circuits on telephone circuits are increasing in magnitude.
grounded neutral electric power lines, and the relation of these As a start in looking more closely at the causes and effects
factors to random separation of buried plant. involved in this problem, it is appropriate to state the three
Noise on telephone circuits due to the proximity of power factors so often referred to as the fundamental components of
lines in aerial construction is an old and much rehearsed subject. inductive interference.
Buried construction, on the other hand, is a relatively new 1) Influence: The combined effects of the characteristics of,
source of noise in telephone circuits, mainly because joint and the power being transmitted by, the electric power line,
construction of buried power and telephone plants has been used which create electrostatic and electromagnetic fields.
only on a limited basis until very recent years. The information 2) Coupling: The electrical interrelation of the power and
developed in this paper is intended to help bridge the gap be- telephone circuits.
tween aerial and buried joint use. Information presented applies 3) Susceptiveness: The characteristics of the telephone circuit
to the evaluation of specific cases of induced noise on telephone including its connected apparatus, which determine the extent to
lines, as well as to the general case. which it is adversely affected by inductive fields.
First, it seems appropriate to show why those in the telephone
industry are concerned about power line noise influence. That is INFLUENCE
to say: Are not telephone circuits quieter today than they were
years ago? The answer, of course, is yes, they are much quieter; Looking at the influence component briefly, we see that a
however, people have come to expect them to improve. Noise on number of things are constantly changing which affect the
telephone lines is one of the major factors used by the public to magnitude of this influence. These are 1) power line configura-
judge the quality of telephone service. However, quality in the tion, 2) amount of power being transmitted and load balance
minds of our customers is not judged against any absolute between phases, and 3) waveshape of transmitted power.
standard, but rather against what they believe they have a right Power line configuration will not be discussed further here
to expect as compared with other services that are a part of because we are primarily concerned with buried plant; conse-
their everyday life; i.e., it is a comparison of what they think the quently, the configuration is usually limited to cable-generally
state of the art is in relation to their other general standards of speaking, a favorable factor. It should be stated, however, that
living. most of this discussion assumes that the power circuits are the
Power line interference or the noise induced in telephone lines multigrounded neutral type.
by power line influence has been a real and difficult problem Power Loads
since the first days of joint use. Looking back over the years to
see what has been done to cope with this problem, it is quite Undoubtedly, we will all concede that power loads are going
up and will continue to increase, and this is an adverse condition.
Also, power load problems, as the communication engineers see
Paper 69 C 1-PWR, recommended and approved by the Trans- them, are of two distinctly different types. They are exposures to
mission and Distribution Committee of the IEEE Power Group
for presentation at the IEEE Underground Distribution Conference, single-phase power lines and exposures to three-phase power
Anaheim, Calif., May 12-16, 1969. Manuscript submitted December lines. If three-phase power lines are well balanced, i.e., if the
10, 1968; made available for printing October 7, 1969. the three phases, the
The author is with the American Telephone and Telegraph power load is equally distributed among
Company, New York, N. Y. 10007. influence is greatly reduced; however, a three-phase line carrying
276 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, FEBRUARY 1970

a large well-balanced load may still have a high influence

because of the combining effect of the zero-sequence components
(stated another way-the odd triple, third, ninth, fifteenth, etc.,
harmonics on three-phase lines, add in phase). Exposures to s
.ff
three-phase power lines are troublesome in many areas where a E
U)

dynamic growth situation exists. Both industrial and residential -C

'R
growths often present a difficult situation for power company L'i
u
z

engineers attempting to keep the loads balanced among the three r')
w
a-
phases. Of course, the greater the unbalance, the higher the 2
-j

influence will be on the telephone circuits. Z)

In rural areas, particularly, telephone companies are often 2

confronted with long single-phase lines that present a high

influence for the following reasons. 100
1) The effect varies directly with the length of the exposure of FREQUENCY (Hz)
the telephone line to the power line, and all the current and Fig. 1.
voltage are "residual" as compared to the three-phase line.
2) To maintain satisfactory secondary voltages, the primary
power line voltage is often increased, by say 10 percent, and the TABLE I
result is overexcited transformers. MUTUAL IMPEDANCE (Q/MI)
The effect of reason 1) is very obvious. The effect of reason 2)
is increased harmonics, or stated another way, the telephone Buried Random Frequency (Hz)
influence factor (TIF) is greatly increased. However, noise due Separation Sites 60 180 540
to this effect will fluctuate with load and will usually decrease Measured values
as the power load increases because as the load goes up the Location A 0.76 0.97 1.05
primary voltage of the transformer goes down, and hence the Location B 0.39 0.46 0.53
Location C 0.41 0.95
transformer is less overexcited. Present-day design and manu- Location D 0.68 0.41
facturing techniques are generally creating a power transformer Calculated values
that exhibits a TIF three times as high at 10-percent over- for 2-inch separation 0.78 1.01 0.92
excitation as compared to its TIF at its rated voltage.
* Assumed earth resistivity 100 t2 m.
Power Current Waveshape
The waveshape of the transmitted electric power is another line harmonic than 180 Hz), the mutual impedance of the joint
constantly changing factor. Although direct comparisons of use aerial plant is 0.78 Q/mi, for the buried plant at 12
present and previous electric power waveshapes are not gen- inches separation it is about 0.57 Q/mi, and at random separa-
erally available, the trend toward the large-scale use of solid- tion it is about 0.92 Q/mi. The computed values for buried plant
state rectifying devices is not comforting to the telephone noise random separation have been verified by tests on physical
engineer. With their rapid switching characteristics, these plant. In Table I, the measured mutual impedance at three
devices create more energy at the higher frequencies. Further, frequencies is shown for measurements made at different loca-
the versatility and favorable low cost of these solid-state devices tions where power and telephone facilities were buried in a joint
are creating new and additional uses of rectifying circuits. trench using "random separation."
As noted in Table I, the bottom line shows the calculated
COUPLING values of mutual impedance for random separation between
buried power and telephone cables. The calculations assumed an
Returning now to the second component in the overall inter- average separation of 2 inches (5.08 cm) between the power and
ference problem-coupling-we find that this characteristic of telephone circuits as typical for random separation. It is assumed
random separation between buried power and telephone facilities that the difference between the measured values and the calcu-
presents an undesirable change from previous joint use situa- lated value for a 2-inch separation is due to the actual existing
tions. A study of the theory of coupling quickly demonstrates conditions at each location; i.e., apparently the average separa-
that mutual impedance is a useful way of handling coupling, tion at location A was about 2 inches and at location B it was
mathematically. Any increase in the mutual impedance between more than 2 inches. Earth resistivity, of course, would also
the disturbing and disturbed circuits represents an increase in account for some of the difference between values at one location
the coupling. and another, as well as between calculated and measured values.
Fig. 1 is a comparison of the mutual impedance of a joint
aerial plant (power wires and telephone cable on the same pole
line) with the mutual impedance of joint buried plant at both Effects of Earth Resistivity
12 inches separation and at random separation. The calculated effects of earth resistivity at 12 inches separa-
The curves in Fig. 1 were computed using Carson's equations tion are shown in Fig. 2 and at random separation in Fig. 3.
[4]. At 180 Hz (one of the more prominent interfering power line Also, the calculated effect of earth resistivity for aerial joint use is
harmonic frequencies) the mutual impedance of the joint use shown in Fig. 4.
aerial plant is about 0.7 Q/mi of exposure, whereas for the In all three figures, variation of mutual impedance with fre-
buried power and telephone plant at 12 inches separation the quency is shown for three different values of earth resistivity.
mutual impedance is about 0.82 Q/mi, and for random separation The effect on coupling of a 100-to-1 change in earth resistivity
about 1.0 Q/mi. At 540 Hz (often a more serious interfering power is not great in any case, but is greater in aerial joint use, where an
WOODLAND: INTERFERENCE ASPECTS OF BURIED POWER AND TELEPHONE LINES 277

F9Ff
I-K-Y 1.-C ,1.6
1.2 - _
z
z
-_
.6
z 1.0 A 10
c: Li C-

:n
BURIED PLANT-RANDOM z
:E
4
A-j BURIEDPLANT- 12 INCHES
F
I c
ltIA - SEPARATION FOR EARTHl |I
.6 BX RESISTIVITY OF
SEPARATION FOR EARTH
RESISTiVITY OF ____
B
1000 METER OHMS-CURVE A
a
=,
5L
. B /;> t /AERIAL JOINT USE FOR
1000 METER OHMS-CURVE Al 100 METER OHMS-CURVE B . B/// 100 METER OHMS-CURVE B
10 METER OHMS-CURVE C 10 METER OHMS-CURVE C
100 METER OHMS-CURVE B I .4
2 | 10 METER OHMS CURVE C '_ .4 f1.
.1 ,F H ~~~FG.2
1: nnn)
.3
10 20 100 1rnr)
l,Wuo 3,UUO 10 20 100 1,000 3,000 10 20 100 1,000 3,000
FREQUENCY (Hz) FREQUENCY (Hz) FREQUENCY (Hz)

Fig. 2. Fig. 3. Fig. 4.

increase in coupling of about 28 percent is indicated for a change will decrease 6 dB for each doubling of the length; i.e., for a
in earth resistivity from 10 to 1000 Q m at 540 Hz. These three manufactured length of 3000 feet (0.914 km) the balance would
sets of curves are used later for evaluation of specific inductive be about 70 dB. The balance will not necessarily decrease as
interference situations. cable lengths are spliced together because of the cancellation
Directing attention now to the third component of inductive effect that random unbalances will have upon one another.
interference -susceptiveness of the telephone plant -it becomes The net effect produces telephone cable pairs with an overall
apparent that three characteristics of the telephone plant are balance (due to capacitance unbalance to ground) ranging
of major importance in evaluating the effect of any inductive from a seldom encountered low of 40-50 dB to a high of 80-90
influence. They are 1) relative interfering effect of different dB and an average around 60 dB.
frequencies, 2) longitudinal balance of the telephone plant, and Evaluation of current telephone cable splicing techniques
3) shielding. indicates that the difference in the resistance of the two wires
Ordinarily, TIF is used as a measure of the relative effect of a cable pair usually does not exceed 1 percent of the single
of different interfering frequencies. However, TIF is a "yard- conductor resistance. Assuming this objective is met, the un-
stick" that includes the effects of coupling and the response of balance due to differences in dc resistance will have about the
telephone subscriber sets and does not include the effect of same effect in converting longitudinal voltages to metallic
unbalance in the telephone plant. For the purpose of this paper, voltages as does the capacitance unbalance of cable pairs.
TIF is not as useful an approach as looking separately at coupling, The equipment connected to either or both ends of a pair of
relative interfering effect, and balance. Coupling as used here telephone wires may disturb the circuit balance considerably
includes the effects of shielding. The C-message weighting curve more than the combined effect of the capacitance and resistance
is a good yardstick for evaluating the relative interfering effect unbalance of the cable pairs. However, this is seldom the case
of different frequencies by this method, and will be referred to except where grounded ringers at the subscriber station are
later in evaluating the overall effect of inductive interference used for identification purposes or for signaling on multiparty
[51-[7]. lines. Many of the noise problems associated with grounded
Telephone transmission on twisted cable pairs inside grounded ringers may be reduced to acceptable levels by using ringer
metal sheaths is not seriously jeopardized by direct electrostatic isolators. Two forms of ringer isolators are currently in common
induction, direct electromagnetic induction, or longitudinal use: cold-cathode gas tube telephone sets and solid-state ringer
electrostatic induction. At least for most practical purposes, isolators. Botb devices have the net effect of opening the ground
these three classes of induction do not cause significant inter- connection of the ringer during the talking period, thereby
ference problems. However, this is not true regarding longitu- removing this source of unbalance. However both devices also
dinal electromagnetic induction. The latter remains an important have some limitations, and they are not compatible with all
consideration in the case of nonmagnetic sheathed cables con- types of circuits.
taining unbalances between the two wires of a pair, with respect To summarize the situation concerning the balance of the
to ground. Any unbalance to ground will, of course, convert telephone circuits, a recent survey of Bell System plant
longitudinal voltages to metallic voltages. These unbalances are indicates that the balance of subscriber lines is usually 50
created chiefly in the following ways: 1) capacitance unbalance dB or better; i.e., the metallic voltages produced by the un-
inherent in the cable manufacturing process, 2) resistance balance conversion of a longitudinal voltage should usually
unbalance due to imperfect cable splicing, and 3) resistance and be at least 50 dB below the longitudinal voltage to ground,
capacitance unbalance due to telephone equipment (central even with grounded ringers. If the use of grounded ringers
office relays, subscriber sets, etc.) connected to the telephone could be eliminated entirely, it would generally be realistic to
wires. assume an overall circuit balance of 60 dB. However it appears
Current manufacturing processes are providing telephone that the use of grounded ringers will be required for many
cable (for Bell System) that averages about 150-pF difference years to come.
in capacitance to ground (or cable sheath) for the two wires
of a pair in a 1500-foot (0.457-km) length of cable. At 1000 EVALUATION OF INDUCTIVE INTERFERENCE
Hz, this represents a balance of about 76 dB for a 1500-foot Having discussed the various components or factors that
length. For longer lengths of continuous manufacture (not longer control the inductive interference situation encountered in the
lengths obtained by splicing), it is expected that the balance construction of joint buried power and telephone distribution
278 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, FEBRUARY 1970

facilities, the next task is to combine these factors to provide FREQUENCY Hz

an evaluation of the overall effect, and thereby form a basis for

50 lT_100 200 1,000 2,000 5,000
0
recommendations for the guidance of the engineers engaged in V1
-10
designing such facilities.
If the harmonic current carried by the buried power cable is
known, this current or currents (one for each 60-Hz harmonic) =I1 _30-; <f ' X1

may be used along with an appropriate mutual impedance to o

-40 4' C MESSAGE WEIGHTING
-T wF 5
CHARACTERISTIC I
calculate the longitudinal voltage (of the particular harmonic
frequency or frequencies) induced in the telephone cable pairs.
If the harmonic current is not known, it will have to be estimated.
The 1932 joint report on power line waveshape of the Edison Fig. 5.
Electric Institute and the Bell System [81 indicates the relation
of harmonic current to fundamental 60-Hz current at the time
of that survey. It has often been used as a guide for such esti-
mates and may still be the best available source of this infor- TABLE II
mation. If more up-to-date local information is available, it
should, of course, be used instead. Percent
The mutual impedance can be calculated for the particular I, of
Frequency VG Zm Power Line 60-Hz
situation when the size or gauge of the conductors is known, (Hz) (volts) (a/mi) (amperes) Value
along with their resistance, relative physical location, type of 60 28.2 0.78 36 100
cable sheath, and earth resistivity. Fig. 1, 2, 3, or 4 may be 120 3.6 0.96 3.7 10
used to select a mutual impedance if the parameters have a 180 1.1 1.01 1.1 3
satisfactory match with those used to calculate the curves 300 0.32 1.01 0.32 0.9
420 0.16 0.97 0.16 0.45
shown in these four figures, the parameters of which are as 540 0.09 0.92 0.09 0.25
follows: 660 0.063 0.87 0.07 0.19
Fig. 4: No. 4 aluminum power phase wire at 30-foot height; 780 0.053 0.82 0.06 0.17
900 0.047 0.77 0.06 0.17
no. 6 copper multigrounded neutral at 26-foot height; 50 pair, 1260 0.046 0.67 0.07 0.19
24 gauge self-supporting telephone cable at 19-foot height. 1620 0.049 0.59 0.08 0.22
1980 0.051 0.54 0.09 0.25
Figs. 2 and 3: Concentric power cable with no. 2 aluminum
phase wire in center and 10 no. 14 bare tinned copper wires on
outside as grounded neutral; 50 pair, 24 gauge alpeth sheath
telephone cable. When the C-message weighting curve correction is applied to
Fig. 1: Same as for corresponding curve in Figs. 2-4. dBrn, it is customary to use a new unit designation, namely
After calculating the longitudinal voltages VL induced in the dBrnc. The curve shown in Fig. 5 has an ordinate scale in dB
telephone cable pairs for each power line harmonic, these voltages units, and conversion from dBrn to dBrnc at any particular
may be converted to reference noise dBrn' as follows: frequency merely requires subtracting the value shown on the
ordinate scale for the frequency in question. When the ordinate
noise to ground (dBrn) = 20 logio (VL/24.5 X 10-6) (1) scale is given simply as relative interfering effect (as is the case in
[9, Fig. 1 ] ) [5], it is necessary to calculate the value of 20 log of
where 24.5 X 10-6 represents the voltage associated with 10-12 the reciprocal of the value of the relative interfering effect in order
watts in 600 ohms. Noise to ground as measured by the Western to convert to dB. As an example, a frequency of 300 Hz has a
Electric 3-ampere noise measuring set is actually equal to 20 relative interfering effect of 0.16 [9, Fig. 11, [5] and 20 log 1/0.16
log1o (VL/2.45 X 10-3), however, because the reference used in is about 16 dB, which correlates with the value shown in Fig.
the noise measuring set has been shifted by 40 dB for convenience 5 for 300 Hz.
when measuring noise to ground. The 40-dB correction does not If we assume a value of 50 dB balance, we can then express the
apply when measuring noise metallic with the set. calculated value of noise to ground as "noise metallic" by
The noise-to-ground quantity obtained using (1) is a function subtracting 50 dB.
of a voltage ratio. In order to determine the interfering effects of This calculation should be made for each power line harmonic
these voltages, it is now necessary to take into consideration the of any significance. The individual values should then be com-
C-message weighting curve [5] shown in Fig. 5. This curve bined on a power basis to find the overall noise metallic in dBrnc
is based on subjective tests where many people compared the due to the proximity of the power line.
interfering effects of different frequencies transmitted to their The overall noise metallic due to any single power line fre-
ear by a modern telephone receiver; the telephone instruments quency should not exceed 15 dBrnc if we expect to meet the
used in the tests were Western Electric 500 type telephone sets. current noise objectives of 20 dBrnc (Bell System) at the ter-
Since 0 dBrn by definition is at 1000 Hz, the C-message weighting minals of the subscriber's telephone set. The difference between 15
curve compares all other frequencies to 1000 Hz. This is par- and 20 dBrnc takes into consideration the fact, that there are
ticularly convenient since the relative interfering effect peaks or is usually several harmonic currents present in the power line,
maximum at 1000 Hz. All other frequencies in the 0-500-Hz and that there are other sources of noise (crosstalk, central
range will have a lower interfering effect. office switching, amplifier noises of carrier and repeatered lines,
battery supply noise, and other induced voltages) that will
combine with the power line interference on a power basis.
1 By definition, widely used in the telephone industry, dBrn is If the single-frequency noise is to be limited to 15 dBrnc, and
the relation in dB of any noise power to 10-12 watts of 1000-Hz
power. if we assume a balance of 50 dB for the telephone facility, we
 WOODLAND: INTERFERENCE ASPECTS OF BURIED POWER AND TELEPHONE LINES 279

TABLE III

Metallic
I, VG C- dBrnc
Frequency Power Line per VG per Message Weighted Metallic Combined on
(Hz) (amperes) Zm Mile Half-Mile dBrn Weighting dBrnc Balance dBrnc Power Basis
60 53 0.78 41 20.5 118 56 62 50 12 15.5
180 1.8 1.01 1.8 0.9 91 28 63 50 13 16.8
300 4.3 1.01 0.43 0.21 78 17 61 50 11 17.2
420 0.13 0.97 0.13 0.06 68 11 57 50 7 17.3
540 0.04 0.92 0.04 0.02 58 6 52 50 2

can calculate the maximum permissible induced voltage at any RECOMMENDATIONS

particular frequency or frequencies. Table II is a summary of The primary purpose of this paper has been to point out or
such calculations. These calculations include consideration of the emphasize the parameters that need to be considered in order to
relative interfering effect shown in the C-message weighting estimate the inductive interference effects of buried power lines
characteristic, Fig. 5. Column VG is the longitudinal voltage on buried telephone lines. Also, the intent has been to develop a
that would produce 15 dBrnc of noise in a circuit with a balance logical and straightforward approach either to estimating this
of 50 dB. Zm in the third column is the mutual impedance be- effect or to calculating it on the basis of measurements.
tween a power phase wire (with a concentric grounded neutral) In order to make some general recommendations, it is neces-
and a telephone wire (inside a grounded metal cable sheath) sary to assume some power line current waveshape. As men-
of the type typically used in joint buried construction employ- tioned earlier, the most recent survey of power line waveshape
ing random separation. The fourth column, labeled I, Power that appears to be representative of the industry in general is
Line, shows the power line current that would produce VG in the [8]. Using information extracted from [8, Table 321, for light
second column in an exposure 1 mile long. The fifth column is residental loads, Table III has been calculated. On the basis
merely an expression of the power line current at the particular of the information in Table III it is recommended that:
harmonic frequency as a percentage of the fundamental fre- 1) The length of the random separation construction should
quency (60 Hz). not exceed one-half mile (0.8 km) unless a detailed analysis,
It should be kept in mind that any one of the interfering based on actual or reliably estimated power line current wave-
frequencies shown in Table II, if present in the amount in- shape, is made.
dicated, will cause 15 dBrnc of noise in the average telephone In order to maintain administrative control and overall
facility. If two or more frequencies are present, as they usually conditions that will permit exposures up to the limits suggested in
are, they will combine on a power basis. 1), the following additional recommendations are made:
60-Hz Interference 2) Joint random separation buried construction should be
utilized only on the subscriber end (opposite end from the
From the standpoint of personnel safety, 60-Hz voltages on central office) of the telephone distribution plant. Feeder cables
telephone conductors should be limited to 50 volts (rms) to should not utilize joint random separation buried construction.
ground. From the standpoint of telephone central office equip- (Feeder cables in telephone jargon refers to the subscriber
ment, a lower limit may be required occasionally to preclude a cables leaving the central office, and out to the point where these
trouble condition. For example, modern electronic offices ter- cables "multiple" or fan out into small cables to accommodate
minate the subscriber lines in sensing equipment consisting of a distribution terminals.)
60-Hz current detector. This current detector inserts about 3) Interoffice trunk cables should never utilize joint random
18 000 ohms impedance between each wire of the telephone pair separation buried construction.
and ground. The detector operates to indicate a trouble condi-
tion when the combined current from both wires of the telephone
pair to ground is about 3-4 mA. The longitudinal voltage re- ACKNOWLEDGMENT
quired to produce such a current is largely dependent on the Basic work done by W. M. Haynes of South Central Bell
characteristics of the telephone pair, and its connected equip- Telephone Company and E. J. Carola of Bell Telephone Lab-
ment at the far end of the line (away from the central office) oratories, Incorporated, was extremely helpful in using Carson's
and in or beyond the area that is the source of the induced 60- equations and deducing a logical approach for rationalizing
Hz voltage. A large number of grounded ringers on the far various facets of this overall problem. The author is also grateful
end of the line would probably be the worst condition; however, a to several others for their comments on earlier drafts of this
sufficiently long cable beyond the inductive exposure could also paper.
provide a low enough impedance to ground to complete the
longitudinal circuit and result in a 60-Hz current in excess of REFERENCES
3-4 mA. When such a current exists in the electronic central [1] L. H. Sessler and L. D. Cronin, "Random separation-develop-
office "sensor," this test circuit denies service to the connected ment and experience," Proc. 1969 IEEE Underground Distribu-
subscriber line and indicates a trouble condition. tion Conf. (Anaheim, Calif.), pp. 252-259.
However, assuming that the 50-volt safety limit will generally [2] "Buried power and telephone distribution systems-analysis of
primary fault tests and evaluation of experience with random
be controlling, from a 60-Hz viewpoint, it becomes evident separation," Edison Electric Institute and Bell Telephone
that for a single-phase exposure, the power line current multiplied System, Rept.
[3] Engineering Reports, Joint Subcommittee on Development and
by the exposure length should not exceed 64 A. mi for a mutual Research, Edison Electric Institute and the Bell Telephone
impedance of 0.78 ohm. System, Repts. 1-50, 1930-1951.
280 IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, FEBRUARY 1970

[4] J. R. Carson, "Wave propagation in overhead wires with Mr. Stumpf outlined the techniques that the author has in-
ground return," Bell Sys. Tech. J., vol. 5, pp. 539-544, corporated in his paper. However, Mr. Stumpf made no attempt to
October 1926. apply this technique to ac exposures in that he felt that the ac
[5] "The telephone influence factor of supply system voltages and system was too complex to obtain definite exposure situations.
currents," in Engineering Reports, op. cit., Rept. 33, October 24, Instead, a figure of merit called IT factor, the product of total ac
1935.
[6] "Frequency weighting for message circuit noise," in Enginreering power current and current TIF, was used as a criterion for ac side
Reports, op. cit., Rept. 45, July 24, 1941. exposures.
[7] W. C. Ball and C. K. Poarch, "Telephone influence factor The application of the dc techniques of [10] by the author to ac
(TIF) and its measurement," AIEE Trans. (Communication telephone exposures is, of course, quite valid and useful as long as
and Electronics), vol. 79, pp. 659-664, 1960 (January 1961 sec.). the following parameters are known: 1) length and relative location
[8] "Waveshape survey on operating power systems," in Engineer of ac power cable and telephone cable, 2) soil resistivity, 3) mutual
ing Reports, op. cit., Rept. 15, February 3, 1932. impedance between power and telephone cable, and 4) magnitude of
[9] Suppl. to [5], Edison Electric Institute, Publ. 60-68, September zero-sequence harmonic currents circulating in the ac power line.
12, 1960. This fourth item is, of course, the most difficult to obtain. It is
impossible to make a general statement regarding the amplitudes of
zero-sequence harmonics of 60 Hz that one can expect to find on
buried power cables. Hence, the recommendation at the end of the
paper that "The length of the random separation construction
should not exceed one-half mile unless a detailed analysis, based on
actual or reliably estimated power line waveshape, is made" is too
general in nature. Furthermore, the author presents no mathe-
matical analysis or data to back up this conclusion.
Discussion It would seem that the problem of ac power cable harmonic
interference is probably still best defined by the power cable IT
factor. A series of tests should be conducted to determine what
relation IT number has to telephone interference in the case of
joint buried cables. Such data would provide a much better rule of
L. A. Kemnitz (Illinois Bell Telephone Company, Chicago, Ill.): thumb to work with than an absolute maximum allowable joint
Illinois Bell has many miles (approximately 580) of telephone cable exposure of one-half mile. Naturally, the author's exact analysis
buried with power cable at random separation. To date, we do not techniques would still be of great value in specific cases.
know of any noise problems in which the main source of noise
influence was a power cable at random separation. However, we REFERENCES
have decided to seek out some of these cables that are most likely to
have noise induction and make specific noise measurements. [10] F. M. Stumpf, "Induction effects of HVDC transmission
systems on telephone communications," 1965 IEEE Internatl.
Conv. Rec., vol. 13, pt. 9, pp. 77-78.
TABLE IV

Noise to Ground
(dBrn)
C-Message 3 kHz
Measurements Weighting Flat Forrest Woodland, Jr.: The following comments pertain to the
discussion by Mr. Kemnitz. The field measurements made at Roselle,
End of loop to Roselle central office 51 68 Ill., would be meaningful if the actual power load had been deter-
End of loop to control point (ran- mined at the time the noise measurements were made. The state-
dom separation section only) 14 49
ment that the load consisted of numerous single-phase transformers
Control point to central office 49 68
Control point to end of loop (ran- with a total rating of 315 kVA at 7.2 kV does not provide a satis-
dom separation section only) 14 48 factory basis for evaluating this situation. Experience has shown
that a new power distribution plant is usually only loaded to a
fraction of its ultimate capacity. Further, the noise measurements
A loop was found in Roselle, Ill., in which the last 4000 feet of would also have been more meaningful if in addition to measuring
telephone cable was buried at random separation with a 7.2-kV noise to ground, they had measured noise metallic. It should also
lead-covered power cable that fed numerous single-phase trans- be kept in mind that, when making noise-to-ground measurements
formers with a total rating of 315 kVA. A control point at the with the 3-ampere noise measuring set, it is necessary to add 40 to
beginning of the buried joint trench section made it possible to the meter reading to obtain the actual value of dBrn being measured.
open the cable pair and measure the noise in the 4000-foot section. To further analyze the actual situation at Roselle it would of course
Noise-to-ground measurements with a 3-ampere noise measuring be necessary to have more data about the telephone plant, such as
set were made as shown in Table IV. These measurements indicate the overall length and exposure of the telephone cable from the
that the noise influence on the overall loop is quite severe, but the central office to the subscriber, and the subscriber equipment con-
contribution from the section of cable at random separation was nected to the telephone line. Further, measurements made on one
insignificant. cable pair do not provide a sound basis for evaluation.
We plan to make other tests and coordinate them with current The following comments pertain to the discussion by Mr. Abbott.
TIF measurements on the power line. My paper is by no means an application of dc techniques to ac
telephone exposures. The whole problem and the theory upon which
the solution is based are derived from fundamental ac theory.
Manuscript received May 28, 1967. I revised my paper to better describe the basis for the conclusions
drawn at the end of the paper. I feel that the revised paper as it
is being published in this TRANSACTIONS, provides adequate clarifica-
tion of the points raised in Mr. Abbott's comments regarding the
conclusions drawn in the paper and the reasons for not using an
IT factor. I wish to thank Mr. Abbott and others who commented
John R. Abbott (Los Angeles Department of Water and Power, on this paper, and particularly for the constructive criticism re-
Los Angeles, Calif.): This paper is an amplification of [10], written garding the conclusions. It was this criticism that demonstrated to
by F. M. Stumpf of Bell Telephone Laboratories, Inc., Holmdel, me the need to rewrite the end of the paper, in particular, and to add
N. J. Table III.

Manuscript received June 12, 1969. Manuscript received July 9, 1969.