tests on counterpoises have not been sufficiently

neorv
Theory ano | ests
and Tests
complete in one respect or another to permit a ra-
tional verification of any theory; although these
previous tests have provided several salient datum
0r the Counterpoise points for checking purposes. It was therefore
deemed desirable to make a set of basic field tests on
counterpoises which would place the analytic theory
on trial and determine its exactness and rigor. The
Different theories have been advanced as to most favorable location for such tests would be in a
of very high soil resistivity and where the
the behavior oF the counterpoise in the region
water level is at great depth; for under such condi-
protection of power transmission lines tions the counterpoise action is very definite. The
against lightning. In an attempt to learn only objection is the difficulty of securing a low
more about the counterpoise and to ,,check resistance ground cathode ray
for the resistivity oscillograph.
,
the theoriesits behavior
the theories of its behavior, a series of basic sers oObviously, if the ground
can be expected of a counterpoise
is nil, no effect
buried in earth
field tests was conducted. The results of which is already an excellent conductor. Unfortu-
both the
these tests, made onsulated parallel in- b d ctests circumstances
themade
natelywere under which the present
were far from ideal for the purpose of
counte s ad te
sulated counterpoise and the buried coun- studying counterpoise action, because the water
terpoise, are reported in this paper. The level was not far down and the soil resistivity not
data obtained are analyzed, and con- high. Therefore the recorded effects were not as
clusions are drawn. or consistent
pronounced serve as could be desired, but
to verify the validity of the theory.
nevertheless
A lightning stroke of surge impedance Z and volt-
By age wave E striking a transmission line tower having
LyV BEWLEY a ground wire, line wire, counterpoise, and footing
ASCA. Generdl Electric Co. resistance, is shown in Fig. 1. In order to simulate
Pittsfield, Mass. this condition, the test set-up was arranged as shown
in Fig. 2. The portable impulse generator, repre-
senting the cloud, was 3,077 ft away from the coun-
terpoise pole. It was rated 0.0125 ,uf and 1,000 thekv
T
HE counterpoise, consisting of con- on open circuit. The surge was initiated from
cathode-ray oscillograph over a spare line conductor.
ductors connected to transmission line towers so as to
reduce the tower footing impedance, has recently The 3,077 ft of line between the impulse generator
been subjected to considerable study as a means of and the counterpoise pole, representing the surge
securing protection against lightning. The purpose impedance of the lightning stroke, was No. 6 copper
of the field tests described herein was to study the wire supported on 30-ft wood poles with 6-unit
action of a counterpoise in detail; particularly in the insulators.
light of an analytic theory (see "The Counterpoise," The portable cathode-ray oscillograph was situ-
by L. V. Bewley, G.E. Rev., 1934, v. 37, p. 73-81) and ated at the counterpoise pole, Fig. 3, and in telephone
with the object of testing the validity of that theory. communication with the impulse generator. Inci-
Theoretical considerations had pointed to 3 major dentally, both these pieces of equipment were
effects present in the behavior of a counterpoise: veterans of several years previous field tests carried
1. The transient impedance, beginning with the initial or surge on by the lightning arrester engineering department
impedance effect and ending with the final or leakage resistance of the General Electric Co. An independent driven
effect, the time of transition depending upon the ground resistivity ground of 60 ohms was obtained for the oscillograph
and, due to the formation of corona, on the surge voltage. about 20 ft from the pole.
2. The coupling effect with overhead conductors, depending pri- At the counterpoise pole there were 2 overhead No.
marily upon the location of the current images and the direction of 6 copper wires at 8.5-ft horizontal spacing and sup-
the counterpoise.
3. The multi-velocity wave components, particularly the low
velocity component predominating on the counterpoise and govern- LIGHTNING STROKE
ing the time of arrival of reflections.
Various investigators have proposed different E
theories to account for counterpoise behavior, but c E e
there has been lack of agreement both with respect m' ''f s t
to the mechanism of the phenomena and with respect GRUOWE
to the magnitude of the several factors involved. LINE WIRE
And this has led, quite naturally, to very pronounced TOWER
differences in the calculated effects. Previous field
Full text of a paper recommended for publication by the A.l.EE. committee _____ ~
+ 71
on power transmission and distribution, and scbeduled for discussion at tbe Fig. 1. Condi- PARALLEL COUNTERPOISE
A.I.E.E. Pacific Coast convrention, Salt Lake City, Utah, Sept. 3-7, i934. tions at RS
tke
Manuscript in pamphletMay
Not publishedsubmitted 31, 1934; released for publication June 25, 1934.
form. srce
stice toeR
oe

AUGUST 1934 1163

ported with pedestal insulators on 30-ft wood poles. The average results of these tests are summarized
One of these conductors could be used to represent a in Table I, the figures in parentheses being calculated
ground wire and the other the line wire. This was values, and the F's coupling factors.
preferable to a permanent ground wire, because it The surge impedance of the counterpoise was
allowed the ground wire effect to be segregated. measured in 3 different ways:
The counterpoise was a 0.25-in. steel wire with a 1 Bytheratioofvoltagesatthejunctionbetweenlineandcounter-
total length of 927 ft. It was initially supported on poise.
short treated wood insulating stakes so as to avoid
the leakage effect, but was finally buried in the earth va.By reflection from the end of the counterpoise grounded through
to a depth of 12 in. It was first tested perpendicular 3. By the ratio of the voltage to the current taken at the entrance
and then parallel to the overhead conductors. to the counterpoise.
When laid perpendicular to the line it was verified
that there was no coupling with the line conductors. In general the 3 methods gave results of the same
order, but the first method is too sensitive to small
TESTS ON PARALLEL INSULATED COUNTERPOISE variations in the applied surges; while the second
method does not take into consideration that the
A 925-ft counterpoise was laid parallel and directly impulse resistance of a driven ground is somewhat
underneath the 2 line conductors, and grounded at lower (about 20 per cent) than its d-c resistance.
its far end through pipes driven into the ground. Furthermore, only the last method provides a way of
It was hoped by this procedure to obtain very defi- showing the variation of the surge impedance as
nite values for the velocity of propagation and surge function of time. From the oscillograms of condi-
impedance and hence definitely locate the voltage tion II of Fig. 4 it is seen that after 0.5 ,sec the
and current images. However, the tests under this counterpoise voltage is increasing, while its current
condition were somewhat disappointing in these is decreasing, so that the surge impedance is in-
respects; because the earth resistivity was so low creasing with the passage of time. There was an
that normal test errors in the determination of the 8.8-ohm resistance connected in series for making
velocity and surge impedance were sufficient to current measurements and therefore the surge im-
cause a wide variation in the values obtained. pedance was
The tests made on the parallel insulated counter- Z, = 398 ohms at t = 0.5 /Asec
poise are shown in Fig. 4. Three conditions were =$3 445 ohms at t = 2.0Asec
investigated: The velocity determination, as ascertained by the
I. Counterpoise present but disconnected, and the surge applied to time required for reflections to return from the end of
one of the overhead wires. This condition is similar to that of
lightning striking a ground wire at midspan. For the purpose of the counterpoise, gave values from 81 per cent to 92
per cent of the velocity of light. This, of course, was
these tests this connection allowed the effect of the ground wire to be
segregated. the average velocity, and should be associated with
II. Surge applied to the counterpoise alone, with the 2 overhead the average surge impedance. The velocity on the
wires isolated. This test segregates the effect of the counterpoise overhead lines was between 96 per cent and 99 per
alone. cent of the velocity of light.
III. Surge applied jointly to an overhead wire and the counterpoise.
This condition yields the combined effect of a ground wire and
counterpoise, and corresponds to the practical case of lightning
striking a tower having both ground wire and counterpoise.

Table I-Tests on Parallel Insulated Counterpoise

Condition I-Fig. 4 II-Fig. 4 I11-Fig. 4

e ....85)......4. ...........0.....3,015 (2,853) ..

......11,520 (11,320)......3,015 (2,850) ......8,7205 (9,200)
e2
32
.......

......2,950 (2,330).......44,400 ..32,400...

86.... (85)................... 6.5 (4.2)
is . ................. 109 (106.5) .......... 80'(74.3)

F,.02 .0242 (0.238). ._..

F". . 0.062 (0.049) . . .......................... ..........

F..,.. .0. 068 (0. 064)

F13.2.......................... 0.269 (0.284)

3,077FT (16-30FT POLES) ------3.368FT (14-30FT POLES)

INITIATING CIRCUIT C2-LINE WIRE

-V(6)52 IN. X 1IN. 1| '+a C
2$
1t' 4 INSUL. el-GROUND WIRE
kNO.6CU.WIRE /< 17S-5
SURGE N 23-KV PEDESTAL INSUL. I
= GEN. CR 3OFT
3-COUNTERPOISE
.4- E G50- (4IN.STEEL WIRE)
925 FT

Fig. 2. Test arrangement Fig. 3. Cathode ray oscillograph at location

1164 ELECTRICAL ENGINEERING,

 Theoretically, the inductance and capacitance can
be computed from the surge impedance and the
velocity, and therefrom the location of the voltage
and current images.
2=2 o
surge impedan
=
s ce
Ziea =

L = relogo)p velocity
v 18 X 10" log
in which
2if) = Zv
2h distance to current image
=
2H distance to voltage image
=
r radius of conductor
=
Z= surge impedance
v = velocity of propagation
Small variations in the measured values of either
Zor v will cause radical changes in the calculated
value of h. Thus if Z/v = 18 X 10-s, then an error
of only 5 per cent in the determination of this ratio
will make a 300 per cent variation in the location of
the current image. On the other hand, inductance
coefficients are not sensitive to variations in h and
therefore its precise determination is not mandatory.
Using an average value of 420 for the surge imped-
ance and 85 per cent for the mean velocity gives
Ls Using X3420Xc 19 and164 =2f,
1220 h
anddesignag2h5
.2h =38 ft
420 X 0.85 X 3(2 1010= 107 X 1011 = 18 X 1011 log
2H =4ft
Incidentally, these figures agree with other data,
as well as theory, showing that the zero potential
plane for voltage images is very close to the surface
of the earth. The capacitance of the counterpoise
as measured by a bridge was 3.13 X 10+ for 925 ft
as compared with 2.64 X 10-9 by the above calcula-
tion.
Using
h= 3 = 19 andH=3 = 2 ft, and designating the
ground wire, line wire, and counterpoise by sub- Fig. 4. Oscillograms taken on insulated counter-
scrips 1, 2, and 3, respectively, there are: poise

r = 0.081 in .. h = 1ft .. L, = 9.55 X (2 X 10-9). . Ki = 0.1154 (18 X 10")

=2= 0.081in8 . =48ft . Ln = 9.55X (2 X 10) .K,, = 0.1154 (18 X 1011)
=3 = 0.125 in .. h = -1ft . L33 = 8.20 X (2 X 10n) . ssKa = 0.1700÷ (18 X 10"n)
I= 0. 081 in .. H = 31 ft. L, = 2.43 X (2 X 10-0) .K/6theKi = -0.0252 - (18 X 10")
R2= 0.081 in . H, = 31 ft .. Li = 0.84X (2 Ya 10-9).K13 = -0.0020 + (18 X 10"e)
R= 0. 125 in....... H, =2 ft .......Ls= 0.84 X (2 X 10-1)....... Ku = -0.0020÷- (18 X( 10"1)

Referring to Appendix A, the auxiliary constants The current surges are

are: il= (0. 1403 f, + 0.0892 f2 + 0.0706 f,)/60
III = I22 = 1.040 al, = 1 Y 0.1407/60
= is= (0.1403 f, + 0. 0892 f, + 0. 0706 f,)/60 (B)
112 In,= 0.038
= a,, = 1 Y12 0.0892/60
= is = (-0. 0612 f2 + 0.4180 f,) /60
I,3 I2,= 0.118
= a,, = 1 YI3 0.0706/60
=
1 =I2= 0.059 a,,
a,,
= -1 21', = -0.1407/60The magnitudes and signs of fl, f2, and f3depend upon
in = 1. 391 = 1 2 = 0.0892/60 the terminal conditions. Corresponding to the 3
v=
1.=0(-910
1.000
VI=0.936
a,, = 1
= 2.6
a23 =
a,, -0.343
Y,, ==
Y23
Y,, =
0.00/0conditions of test, there are:
0.0706/60
-0.0612/60
a,, = 2.965 Y,, = 0. 418/60 CONDITION I
In terms of the multi-velocity wave components
fl(x- vlt), f2(X - V2t), and f3(x - v3t) the voltage
is == 0 = vurrent in neewire
surges on the 3 conductors are i, o current in counterpoise
e, = f, + f2 + f Hence by eqs A andB
e2 = - f. + f2 + f3 5. (A)
e, = - 0.343f2 + 2.97f, f, = 0.381 Eo; j2 = 0.540 Eo; f3 = 0.079 Eo

AUGUST 1934 1165

and therefore, writing Eo(vi) for a component of and therefore,
magnitude Eo traveling at a velocity vy, there is by el = -0.244 Eo(v2) + 0.308 Eo(v3) = (0.064 Eo)
eq A: e2= -0.244 Eo(v2) + 0.308 Eo(v,) = (0.064 Eo)
el = 0.381 Eo(v1) + 0.540 Eo(V2) + 0.079 Eo(vs) = (1.000 Eo) e3= +0.084 Eo(V2) -1 0.916 Eo(v3) = (1.000 Eo)
e2 = -0.381 Eo(v,) + 0.540 Eo(v2) + 0.079 Eo(vs) = (0.238 Eo) il= [-0.0218 Eo(v2) + 0.0218 Eo(vs)]/60 = 0
e3 = - 0.185 Eo(v2) + 0.234 Eo(vs) = (0.049 Eo) i2= [-0.0218 Eo(V2) + 0.0218 Eo(v3) ]/60 = 0
Due to symmetry of the overhead conductors with is3 = [ 0.0149 Eo(v2) + 0.1289 Eo(v3)1/60 = 0.1438 Eo/60
respect to the counterpoise, there is no v1 component Thus the initial coupling factor is about 6.4 per cent,
in the counterpoise surge, and the terms of this and the surge impedance of the counterpoise is
velocity balance between el and e2. The coupling e_ 60 417 h
factors are is 0.1438 oms
F1.,= 2 = 0.238 and F.
el el
= 0.049
is By =-3
the approximate method of Appendix B, there
for comparison with the values (in parentheses)
The currents are, by eq B: given by the complete multi-velocity theory:
il = [0.0536 Eo(vi) + 0.0482 Eo(v2) + 0.0056 Eo(v3)]/60 = V = = L = 5
2=NPlLl=N9-195
9 = 0.978
.7 (0.981)
091
(0.1074Eo/60)
i2 = [-0.0536 Eo(vi) + 0.0482 Eo(v2) + 0.0056 Eo(vs) 1/60 = 0 V3 = p33/L33 = V5.95/8.20 = 0.852 (0.836)
i = [ -0.0330 Eo(v2) + 0.0330 Eo(v3) 1/60 = 0 Z= Z2 = v = 560 (558)
The surge impedance of the ground wire, therefore, is zs = -3 = 419 (417)
Z= ei/ii = 558 ohms e2 L23 0.84
es L33 (1 + v/lvI) 8.20 (1 + 0.852/0.978) 0° (0.06)
CONDITION II
The agreement is seen to be quite good, and thus
il = 0 = current in ground wire justifies the use of the more simple approximate
i2 = 0 = current in line wire method.
es = Eo = voltage applied to counterpoise
Hence by eqs A and B CONDITION III
fi = 0; f2 = -0.244 EO; f3 = 0.308 E
Referring to eqs 1 to 10 of Appendix A there results
fi = 0.358 Eo, f2 = 0.273 Eo, f, = 0.369 Eo
~~~~~~~~~30
120 and the voltage and current equations are
- - r -

000 IOC-
- _ _ _
00 -00_-,t%_ -
- -
-,-,:00 ei = 0.358 Eo(v1) + 0.273 Eo(v2) + 0.369 Eo(v3)
= -0.358 Eo(vi)
+ 0.273 Eo(v2)
+ 0.369 = (0.284 Eo)
(1.000
Eo(v3)
=

Eo)
900 98o <e-s0.093
0 Eo(v2) + 1.093 Eo(v3) = (1.000 Eo)
700 70 70._
600 6C - - - 0 _6 Fig. 5 (left). Transient impedance of 925-ft buried counterpoise, from oscillo-
50050. . .
..V0 grams of voltage and current

32002C . .0
_L.EAKAGE _ Fig. 6 (below, left). Transient current of
impedance
voltage and 200-ft buried counterpoise, from
oscillograms
100
0o I° 0
0

0 2
MICROSECONDS
3
RESIST4iANCE 1
RESSTNC

5
10

Fig. 7 (below, middle). Test points and curve of uniform leakage resistance
Fig. 8 (below, right). Method of determining velocity of propagation on the
counterpoise
2
I 130
1-
~~~~~~~~~~~~~~~~~~~~~~~~~~
120 12 l 120
fr 11011011
110 hLINE WIRE
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
90 9
Lu 0
707 X iAC 0 0010 9 X 9020 <0Q
9~~~~~~~~~~~~~~~~~~~~~~
f V W WAVE
L70F
SLOW

Q-
COUNTERPOISE~~~~~~~~~~~~~~~OUTEPOS

60 C ~ - 2 O 20 40 60 0 00 2 MCOEO

MICRSENDS LENGTH OF COUNTERPOISE (fEET)

1166 ELECTRICAL ENGINEERING

h = 1 0.0503 Eo(vi) + 0.0244 Eo(v2) + 0.0261 Eo(v3)1/60 =
(0.1008 Eo/60)
i= [- 0.0503 Eo(vi) + 0.0244 Eo(v2) + 0.0261 Eo(v3)]/60
(0.0000 E0/60)
is = [ - 0.0167 Eo(v2) + 0.1543 Eo(vs) 1/60 = (0.1376 Eo/60)
In this case the coupling is 4 per cent higher than
with the counterpoise disconnected; and due to the
mutual coupling between the ground wire and coun-
terpoise the currents are less than under conditions I
and II. It will be observed that when the counter-
poise is connected, the surge which it carries is essen-
tially a low velocity wave,
TESTS WITH BURIED COUNTERPOISE
The counterpoise was buried to a depth of about
12 in., parallel and directly underneath the line
conductors. Tests were made on lengths of 200, 500,
and 925 ft with surges of 15 kv and 90 kv. The
latter was the highest voltage at which current
measurements could be made without making special
arrangements to insulate the portable cathode ray
oscillograph truck. For slightly higher voltages the
truck tires would arcover.
In Figs. 5 and 6 are replots of voltage and current
oscillograms taken on the 925- and 200-ft counter-
poises, respectively, with the overhead conductors
isolated. The ratio of voltage to current here is
plotted as the transient impedance. Initially the
counterpoise behaves as a surge impedance and
finally as a distributed leakage resistance to ground.
The equation for the transient impedance is given in
Appendix C. The points on the plotted impedance
curve lie on a smooth curve beyond 0.5 pssec, but the
accuracy is very poor between 0 and 0.5 /sec. How-
ever, the extrapolation of the impedance curves on
these and other plots indicates that the initial, or
surge impedance, value lies between 120 and 160
ohms. From this value the transient impedance
falls in a somewhat exponential fashion until the
final, or leakage resistance, value is reached. In the
case of the 925-ft counterpoise this final value has not
yet been reached at the end of 6 ,usec (at which time
reflections from the generator interfered). But on
the 200-ft counterpoise the transient impedance has
very definitely reached its steady state value at the
end of 1.5 ,usec. At the instant of maximum voltage
the transient impedance is 88 and 112 ohms, respec-
tively, in Fig. 5 and Fig. 6. Very likely this impedance
is considerably less at the much higher voltages
corresponding to natural lightning. Incidentally,
in each of 5 plots which have been made, the imped-
ance curve closely parallels the tail of the voltage
wave. The leakage resistance, Fig. 7, measured
77.5, 31.5, and 14 ohms on the 200, 500, and 925 ft
lengths of counterpoise, respectively.
The propagation of a surge in a counterpoise buried
in low resistive earth appears to be as much diffusion
as traveling wave phenomenon. For this reason
attempts to discover a velocity of propagation from
reflections on the oscillograms proved futile. The
abrupt change at 1.5 ,u~sec on the tail of the voltage
wvave for the 200-ft counterpoise, Fig. 6, as con-
trasted with that of the 925-ft counterpoise, Fig. 5,
was at first taken for the return of a reflection.
AUGUST 1934
In order to deternmine more carefully the velocity,
E. J. Wade devised amost ingenious method, the essen-
tial elements of which are shown in Fig. 8, the waves
being replots of actual oscillograms. The surge
was applied to both the line wire and the coun-
terpoise. The cathode ray oscillograph was placed
200 ft farther along where the counterpoise was
opened as shown in Fig. 8, and a series resistance of
8.8 ohms inserted. By closing a the fast wave on
the line was recorded. By closing b (with a open)
the flattened front slow wave on the counterpoise
was recorded. Finally by closing both a and b the 2
waves superimposed after a time at which could be
determined by placing the oscillogram of the fast
wave over that of the combined wave so that their
fronts coincide, and observing the point at which
they diverge. It will be appreciated that the success
of this method depends upon the 3 resistance ele-
ments being of widely different values, and the one
to the line wire should be high. In Fig. 8 the slow
wave arrived 0.3 ,usec after the fast wave, and the
velocity therefore is (since the fast wave travels at
the velocity of light):
I I 1 200
v3 = -3 = +
-1 A = A + == 0 + 20 = 400 ft/j.sec
But when this test was repeated at 500 ft the ve-
locity was apparently only 330 ft/ysec. The reason
for the decrease in velocity as the distance traveled
increases is given later under "Discussion." The
method of testing determines only the average
velocity. Now the longer the length of counterpoise
under test the less the influence of the initial higher
velocity and the more nearly will the average ve-
locity be equal to the final velocity. On the basis of
this argument the velocity will be assumed as 30 per
cent that of light. Taking the surge impedance as
Z3 = 142 and the velocity at V3 = 30 per cent, the
inductance and capacitance coefficients, as well as the
position of the equivalent current images, are as
follows:
2 1 2hg Z3 142 15.8
L = 10'9 r V3 0.3 X 3 X 1010= 109
2700
2h 8 X 1228 ft
=
1 1
C = 1 x 10" log (2H/r) - Z v
1 0.80
142 X 0.3 X 3 x 1o1 = 1012 farad/Cm
.2H = 0.25 in.
In earth of such low resistivity, burying the
counterpoise may raise the effective position of the
ground current considerably, because current is then
flowing above and on both sides of the counterpoise
as well as below it. In such cases the depths of
current as determined for the insulated and buried
counterpoise are not comparable. But in high
resistivity soil, where counterpoises are most likely
to be used, there will probably be little difference
found in the location of the current image for either
an insulated or buried counterpoise.
Current distribultion along the counterpoise was
obtained with the cathode ray oscillograph at the 0,
200, and 500-ft points. Representative oscillograms
1167
 Table 11-Tests on Buried Counterpoise low voltage surge of 15 kv and Fig. 12 is for a 90-kv
surge. The data (averaged) is summarized in
Fig. 11 Fig. 12 Table II.
Condition II III II m By eq 27 of Appendix B the mutual impedance
between the counterpoise and the overhead con-
-Ci. 1,600.. 14,480 . 5,990.. 90,150 ductor, based upon the current image at 28-ft depth,
e2 1,600. 4,050. 5,990 .. 23,820
s1
..

.316,400 14,480 .. 94,500 ..

. 90,150
.4540... 187
i3
178 . 161.1 ... ,1071,107 V/2V3 1.29 /1 X 0.3
FP.. 0.098. 0.063
2.028..0.264//
P13,
ZJ.2=L3 V2 + Vs 3) 3X101.= 9.0
From Fig. 11 the current is 178 amp. Therefore,
are replotted in Fig. 9. The initial current wave is the induced voltage is
quite steep, having a front of less than 0.5 psec. e2 = Z3.2i3 = 9.0 X 178 = 1.6 kv (1.6 test)
There is considerable flattening of the front and drop Likewise from Fig. 12 the induced voltage is
in magnitude as the current wave travels along the
counterpoise. The average values obtained from e, = 9.0 x 1100 = 9.9kv (6test)
these tests at 2.5 ,usec were: This latter is a very poor check. Presumably the
Distance 0 ft 200 ft 500 ft only difference in the 2 tests was the higher voltage
Current 134 amp 107.6 amp 14.4 amp in the case of Fig. 12. But if this made any differ-
ence at all one would expect it to increase rather than
Since the counterpoise current is so rapidly drained decrease the coupling. The author has no explana-
off, the coupling effect with the overhead conductors tion to offer.
practically vanishes within a short distance-say
beyond 300 ft. Attempts also were made to measure DISCUSSION
the currents by surge crest ammeters, but the re-
versals due to reflections were too large to permit In seeking evidence from these tests in support of
accurate results. the multi-velocity theory, perhaps the most con-
The superposition of voltage waves taken on clusive is that shown in Fig. 4. According to the
counterpoises of 200, 500, and 925-ft length is shown conventional theory of traveling waves there is no
in Fig. 10. There is obviously nothing to gain current on an isolated conductor-it merely floats in
under these conditions by using a counterpoise the electrostatic field of adjacent conductors and
longer than 200 ft for applied wave fronts less than takes on the corresponding potential, and therefore,
1.5 psec. In soil of higher resistivity a longer length the shape of an induced surge is an exact replica of

15C Fig. 9(left). Current waves in 925-ft "3.-.

130 - - -- ..~FT
140.counterpoise at 0, 200, and 500-ft
pont
. -

_L
-v_-20F_- a- <ig.
FtXT
- 10) (righlt). Superpositioan of volt- vA||||S tWItw
-F i age waves for pountepois of different t
50-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
30.500 Fr MICROSE~~~~~~~~~~~~~~~~~~~~~~~~~~0COUNDS POS

Io -lnt hs ]lu gsre u codn otemli

Fig.is(rig th inducing sur.

ishont accordion to the mnsulati

300 ft are worth the extra cost. On the other hand,
very short counterpoises of the order of 50 ft may
give rise to substantial reflections which would raise
the tower voltage.
In Figs. 11 and 12 are shown the effects of the
buried counterpoise on coupling for the 2conditions:
(II) Surge on the counterpoise alone with the over-
head conductors isolated; and (III) surge on both
the counterpoise and ground wire. Figure 11 is for a
counterpoise the voltage surge is
er= -0.185 Eo(vr) + 0.234 Eo(v) = (0.049 Eo)
-When this surge reaches the end of the counter-
poise, which was grounded through driven pipe
grounds, it reflects negatively and returns to the
starting end of the counterpoise. The fast wave by
thlat time has separated from its slower companion
and so appears on Fig. (41) as a positive bump.

1168 ELECTRICAL ENGINEERING

 .. ~~~~TRAVELING WAVE
~~~~~~~~~.~~~

(a) (b) (C)

GROUND SURFACE

*-ZERO POTENTIAL PLANE

RSNTOSURFACE
XliiOWNW=RTREND OF CURRENT

Ga ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

o ~~~~~~~~~~~~~~~t-t
Fig. 1 3 (above). Phenomenon at the
wave front

u~~~~~~~~~~~~~~~~~ "IJL I-RELs Top. Equivalent network at successive in-

stants
Middle. Current and zero potential plane
in the earth
Bottom. Variation of surge impedance and
velocity

buried (leFt). Oscillograms

Fig. 11 counterpoise taken on
for a tow voltage
* ~~~~~~~~~~~~surge
oJ'1 5 kv
Fig. 12 (center). Oscillograms taken
on buried counterpoise, For a 90-kv
surge
But the slower wave is the larger of the 2, and there- Under condition III the bump in e2 is less than
fore, dominates upon its arrival, reducing e3 almost to under condition II. This is because there are 2
zero. inducing currents involved, il and i3, and due to
Again in Fig. (411) the positive bump is evident on reflection i, decreased when i3 increased, but not so
e1 and e2, whose initial components were found to be: much, so that the net effect on e2 is still a positive
el = -0.244 Eo(v2) + 0.308 Eo(v3) = (0.064 Eo) bump.
e2 -0.244 Eo(v2) + 0.308 Eo(v3) = (0.064 Eo)
= The multi-velocity components, therefore, are
Undercondition . evidentthattheinseen to be present and behaving in conformity with
Under conditio sIIe itt iss evident that the mduced theory. The numerical checks are probably as good
waves el and e2 register a positive bump at the same as can be hoped for in this kind of calculation; es-
time when the main surge e3 experiences a voltage pecially in a situation where the order of magnitude of
drop and i3, a current increase. Thos cs rather con- some of the quantities involved is not very much
clusive evidence that the isolated conductors are not greater than the order of magnitude of the test
merely "floating in the electrostatic field of the main errors.
surge," but are conforming to the main surge current. In many problems dealing with traveling waves it
This is all in complete agreemetnt with multi-ye- is permissible to ignore the phenomenon at and
locity theory, which predicted that the induced volt- immediately following the wave front, but in study-
age from a counterpoise is due to the counterpoise ing the action of a counterpoise some cognizance
current. Indeed, according to Appendix B should be taken of this behavior, since the major
e2 L23 effect is confined to the first few microseconds.
ea L3a (1 + V2/v,) When the surge impedance for the insulated counter-
Thus the coupling factor is constrained by the limits poise was plotted as function of time by taking the
instantaneous ratios of voltage to current from
i (L2a)
2 L3as
\ (.~'\ e2 ) g (Loscillograms with fastsharply was observed that the
sweep itfrom
ea L3 impedance dropped a high to a mini-
or has no more than a 2:1 range as far as capacitance mum value within the first half microsecond, then
coefficients (which are included in the velocities) are increased slowly as shown in the bottom sketch of
concerned. Fig. 13. It has already been remarked in this paper

AUGUST 1934 1169

that the average observed velocity decreases with the multiplicity of counterpoises of this length arranged radially. Very
length of the counterpoise. In order to explain short counterpoises give rise to be
reflections and are oftherefore not so
efficient as one long enough to practically free this source of
these changes in surge impedance and velocity the additional voltage.
situation of Fig. 13 is envisaged. When the abrupt 8. The multi-velocity theory appears to be adequate for numerical
wave front first arrives at a given point the current calculations as well as for a description of the mechanism of counter-
distribution in the ground is governed by the ca- poise action. But for routine design calculations a much simpler
pacitance of the earth, Fig. 13a, and thus the zero
potential plane with respect to voltage images is at
some depth below the surface. Consequently, since
A
ppe
ndix A
Z=-V7, and V- 1z/V7and N/-
the effective
NI ~adV adte
= fetieC C The multi-velocity wave components on a 3-conductor system
is small on account of the depth of the zero potential comprising a ground wire, line wire, and counterpoise are (see
plane, the surge impedance is high and the velocity "The Counterpoise," by L. V. Bewley, G. E. Rev., 1934, v. 37, p.
is high. These starting values are designated as 73-81):
points a in Fig. 13. There then ensues a very fast e1 = ailfi (x - vit) + al2f2 (X - V2t) + alafs (x - V3t)
transient at the end of which the resistance network e2 = a2lfl (x - vlt) + a22f2 (x - V2t) a23f3 (X - v3t) J (1)
dominates over the capacitance network, and the e3 = aiifi (x - vit) + a32f2 (x - V2t) + a33f3 (x - V3t)
zero potential plane rapidly rises to the surface of the il = Yllfl (x - vlt) + Y12f2 (X - V2t) + YlJ3f (X - v3t) )
earth, while the current continues its relatively slow i2 = Y21fl (X - Vlt) + Y232f2 (X - V2t) + Y233f3 (x - v3t) (2)
downward trend. Thus the end of this readjust-
ment finds the effective C much increased, but with in which the velocities v,, v2, v3 are given by the roots of
no material change in the effective L. Therefore, (Ill- v-2) (122 - V-2) (133 - -2) + I12123I13 + 113132121-
the surge impedance reduces to a minimulm and the (Ill - V-2) I23I32 - (122 - v2) 113131 - (U33 - v-2) I12I21 = 0 (3)
velocity to a lower value during this transition stage, and the a coefficients are (for r = 1, 2, or 3)
designated by points b in Fig. 13. Thereafter the
a a12 - cii - 1 (arbitrarily)
I
effective capacitance does not change, since the zero
potential plane is already at the surface, but the aI=I2113-123(1iVr11 ) I31I23-I2i(I33Vrr) 1
current continues its downward displacement, there- 112123-311(122 -V-2) (122-v,-2) (I33-v-2) -132132
by increasing the effective L, and tchus causing the a312 ++ (4)
surge impedance to increase slowly and the velocity -I2In2121 (I22-V,2 -2) 121132)-3ISv(2) 1321
to decrease slowly. These effects, which are clear 12I23-I13(I22-V,2) (122-V2 (133-V,-)-132I23
and definite for an insulated counterpoise, are masked where, in general,
by the leakage effect in the case of the buried counter- Ire = (L1rK2. + L2rK2, + L3rK3.) (5)
poise. Y= (K,,a,, + K,2a2 + K,,3a3,)v, (6)

CONCLUSIONS The L's and K's being the inductance and capacitance coefficients.
If there are 2 ground wires in a horizontal plane, or 2 parallel
counterpoises, then the inductances and electrostatic coefficients
The The
conditions ulnder which these tests were made
conditions under which these- tests made
were
are to be calculated as for an equivalent single conductor. Thus
were not conducive to the most accurate type of for 2 ground wires 1 and '
determination of the factors involved. The effects 2 2 =n
were not pronounced enough, and the test errors too L,, = -log, self inductance of the pair
erratic, to permit of any "rock-bound" conclusions;
but the following facts seem to have been established. L12 = 2 logf,
laag mutual
c - 0-9 =b
nductance of the
conductor
paor to the line
1. The impedance of a buried counterpoise is a variable starting
at a value equal to its surge impedance and ending at a value equal where
to its leakage resistance. The transition may require from 1 to
10 pjsec, depending upon the length of the counterpoise and the 2 hi = height of ground wire from its image
earth resistivity. Propagation of surges on counterpoises is thus r = radius of ground wire conductor
as much diffusion as traveling wave phenomena. m = separation between ground wires
2. The coupling effect of a counterpoise depends primarily upon b = separation between ground wire 1 and line wire
the position of the current images, and is not greatly dependent upon a = separation between image of 1 and line wire
the capacitance relationships. The higher the earth resistivity and b' = separation between ground wire 1' and line wire
the greater the depth to the water level, the lower the current images a' = separation between image of 1 and line wire
and the higher the coupling.
3. The earth's surface may be considered the zero potential plane Similarly for a pair of counterpoises the electrostatic coefficients are
for voltage images of traveling waves. likewise
4. Multi-velocity waves exist on a counterpoise, but the only 2 Hi
one of importance is very slow, traveling at about 1/3 the velocity of Pi, = 18 X 1011 log,
light. The arrival of reflections may be estimated on that basis.ai
5. A new and accurate method for determining velocities on coun- I aa'
terpoises has been devised (E. J. Wade). This method indicated P12 = 18 X 10"l logE -¶j bb
that the initial velocity of the wave when it first enters the counter-
poise is higher than its ultimate velocity. This is reasonable where 2 H1 is distance to the electrostatic image and R is the radius
theoretically. ~~~~~~~~~~~of
the effective corona envelope.
6. Due to the leakage the current in a counterpoise drops off rapidly If a lightning stroke of surge impedance Z and incident wave E
and the front of the advancing wave is greatly flattened. For this strikes a tower having ground wire and counterpoise, and tower
reason the coupling effect vanishes at points a few hundred feet from footing resistance R, then
the tower.
7. Counterpoises more than 200 to 300 ft long do not appear fi = E [(a2 - 1) 1Y33- (a3 - 1) Y32] =b1E0 (7)
justifiable. If greater effect is desired it should be secured by a A

1170 ELECTRICAL ENGINEERING

f2 =
Eo
- (a33( Y21-(a31-1) Y23] =
[(a33-1) b2E(
Now disregarding the multi-velocity wave components of Ap-
pendix A, a linear set of equations may be tentatively written
between the voltages and currents on the system:
f$ =
A- [(al
- 1) Y22 - (a32 - 1) Y2'1 = b3Eo (9)
el Zllil + Z12i2 + Z13iS = ground wire
e3e10= Z31il + Z32i2 + Z33i3
= Z21i1 + Z22i2 + Z23i3 = line wire (23)
A = [a32 (Y23 - Y21) + a3l (Y22 - Y23) + a33 (Y21 - Y22)] (10) = counterpoise
Eo= VWhen multi-velocity components do not exist these are the exact
(2E)
1 1 ) (11) canonical equations for traveling waves, but if multi-velocities do
z [I +- z +2(Yll+ Y31)bl+2(yl2+ Y32)b2+2(Y13+ Y33)b3J exist they are only approximate, however give close results for the
first few microseconds. It remains to identify the coefficients.
The equations of this appendix permit the calculation of the behavior Obviously
of a counterpoise. Reflections up and down the tower have been
ignored in accordance with the justifications given in original paper Zii = ;LlipIi = surge impedance of ground wire
on this analysis. ("The Counterpoise," G. E. Rev., 1934, v. 37, p. Z22 = = surge impedance of line wire (24)
73-81). Z33 =
/
Las = surge impedance of counterpoise

PERPENDICULAR COUNTERPOISE Suppose that only current il is flowing. Then

e2 Z21
If the counterpoise is perpendicular, the voltage equations are -= = F12

el =
Y22Eo(x -vilt)-
v1t Y2,Eo(x
1E --v2t) (12) and since both conductors are overhead, F12 must be computed by
( Y22- Y21) eq 19. Therefore
= a2lY22Eo(x-(Y22
vit) - a22Y21Eo(x - V2t)
- Y21) Z21 = Z11F12 = Zll IY22-- a22Y21 Y21 (25)
L Y2- 2
Eo 2E YY\) (14) Likewise, interchanging subscripts
-~+ (Y14 ~~~~~~~~~~~~~~~Z121[2 1a2I-lY2
-a21(6
+ 2R Y33Z + 2 22 222,= Y
Y22- Y21 = Z22F21 = Z22 (26)
y1l y12 -
1 U(1l + 122) /(111 I-122)2 + 4 112121 (15)
2 For the coefficients involving the counterpoise, the simplification of
a2l = (v,-2 - 113)J12 the type of eq 22 can be used, thus:
a22 = (v2-2= (Ku- +Ill)/I12 ( ) = LIS 1 L v1v3
Yi, =
a23K32)v1
(Kil + a21Kl2)V1 )
( )~~~~~~~~~~IS=
Z13 = Z33F31 = Z3 L33 1 +33F3(v3/vl)
Z3 -= 1 \v, + v3
V12 = (K11 + a22K12)v2 L1__
Y21 = (K12 + a2lK22)v1 (17) L23 1 V2VS3
Y22 = (K12 + a22K22)v2 L33 1 + (VS/V2) V2 + v3
Ill =
I12 =
LilK1i + L12K12
LliK12 + L12K22 (18)
L13 1
Z31 = Z11F13 = Z11 , 1 (v1/v3)3 L VLV3
(27)
121 = L22K12 + L12K2l (18) z31 + 23l V + V
122 = L22K22 + L12K12 L23
Z32= Z22F23 = Z22 L 1 + (
1 2V___
L23 v2 V v3,

Appendix B It is seen that Z31 = Z13 and Z32 = Z23.

Having established all the coefficients of eqs 23, these equations
The general analysis given in Appendix A is not only complicated may be used directly.- Assuming as in Appendix A that a lightning
and laborious (a complete set of computations requiring at least 2 stroke of surge impedance Z and incident wave strikes a tower
days' work), but small errors made in the auxiliary constants may having ground wire, line wire and counterpoise,, and tower footing
throw the final results quite far off the correct values. In this resistance R, then remembering that i2 = 0 on the line wire:
appendix a much shorter approximate method is developed which el = Eo = Z11i + Z12i3
should suffice for ordinary calculations. e3 = Eo-Z21i1 + Z33i3 . (28)
Referring to eqs 12 and 13 of Appendix A, the initial coupling e2= Z21i1 + Z23i3 )
factor between 2 conductors No. 1 and No. 2, the main surge being
applied to No. 1 and the induced surge measured on No. 2, is But the total current from the stroke in terms of the incident wave
E and the reflected wave E' is
a2l Y22 - a22 Y21
F12 _(19) Y22- Y21 (19) E -E'
E-jy' 2E
2E E±+El2 _ E,
_ E + E' = 2E E° = 2 2(il + is) + FEo (29)
z-
Now if one of these conductors lies on the ground surface, the Q
mutual capacitance coefficients are zero (K12 = K21 = 0). On this Solving for i2 and i3 from eq 28 and substituting in eq 29 there results
basis, substituting eqs 15, 16, 17, and 18 into 19 and simplifying
(it is quite a long job) there finally results = E Z33-z( = bE
FZ11Z33- Z132 /
F12 =PiL1 + (LsL22
-/L22)
(20)
(20) = Eo Z
Zi - Z13
-2 -z ) = bsEo s (30)
or neglecting L122 in comparison with L11 L22 and substituting F0 = (1 2Z (b 2E b3) + Z/R)

v= L11 L2=FJ2 (21) Hence substituting in eq 28

the initial coupling factor may be written Eo=
3=
cl= 1 + 2-bs)Z/Rl (3
b+

±- g112
L20
=
(L-11I1 + (vl/v2) (22) e2=(Z2b1 + Z23b2)Eo 2E (Z21b1 ±
Z23b2)

AUGUST 1934 1171

Appendix C Toward the Making
In order to approximate the effects of the transition of the im-
pedance of a counterpoise from its initial condition of surge im- of a Prof
ression
pedance to its final condition of leakage resistance, the following (Continued from page 1148)
method of analysis appears to have some merit. Assume the
leakage resistance Rs = l/gl to be-concentrated at the end of a
counterpoise of surge impedance, Zs, and length T (to be defined o phs
am
later). Then successive reflections will eventually supplant Z3 chosen to be the inspirer rather than the policeman
with R3, and the problem reduces to that of computing these re- of the profession. The intangibles of the spirit,
flections. The time T is to be chosen so that the transition occurs the creating of an atmosphere, a tone, a fellowship,
in the required length of time, either as estimated from tests or as have far outweighed the letter of the law. In its
calculated from the following formula:
Z3(t)=
contributions to public welfare, the Institute has
wisely contented itself to work through the most
1 (32) inclusive agencies of the profession, the American
X)2
-- r2-at L Engineering Council, the national and interna-
1)2=2
(2s +4c14 CC - X s
tional standardizing agencies, and most recently
the E.C.P.D.
where
I (2s - 1)21r2 g2 (2s - 1) ?r RELATION OF THE INSTITUTE TO THE
- 2 AJ LCl12 C~2lVT- 91RMODERN STATE
a = g/2C In conclusion, only a hint can be given of the
assuming (g/,QC) to be a small quantity, this reduces to significant contribution of the Institute
growing and evolving idea of a professionto asthea
1(t) functional
call theelement in the complex
gl I e-at
r',
8 cos 04 sin(3'(33)
C we modern state. The oftraditional
social organisms
fabric
gi 1l _ e-astZ[fL(25
's =l1
- 1)27r2 - giVl (2s
(25s- 1)7
1)]rJJ; of society with its political and geographical strands
is breaking apart under the stress of a complex life.
The first term under the summation sign is a triangular wave, and In its place we see emerging the modern conception
the second term a rectangular wave. Therefore it is easily seen that of the corporative state, a society in which the
constituent units are not parties or political sub-
Zs(t) gi
= I at = divisions, but functional groups, such as trade
associations, labor unions, consumers' cooperatives,
IL (34) professional guilds, agricultural groups, and the
Zs(t) = .C
at t 0 like. The idea which underlies much of our de-
mocracy and in which Jefferson and his disciples
Thus if g, L, C of the counterpoise are known, Z3(t) may be computed so firmly believed, that the citizenry acting col-
and the time of transition from V77C to l/gl readily found. The lectively would transcend their wisdom and in-
transition is brought about by the presence of the term e-at. There- telligence as individuals, is sadly discredited. Too
fore considering the transition as essentially complete when E-t = often the reverse is all too evident in political life.
In professional bodies such as our own the idea
3 6Cg t . of collective
group we tendsuperiority
has been
to act on a higher planevindicated;
as a
a of idealism and
Now consider any set of wave components (fi.f2,f3) on the ground
of devotion
of to thesociet
commonapergood than
to asdependi
individuals.
more
wire, line wire, and counterpoise, in accordance with eqs 1 and 2 The future of soclety appearS to depend more
of Appendix A. When these waves reach the end of the counterpoise and more upon the vitality of such functional
they will give rise to reflections (fl', f2',fJ') and transmitted waves groups. Our greatest concern may well be to extend
(if', f2', fs") and the transition point equations become professional ethics and ideals to the wider areas of
ei + ei' el" industrial and public life. The Institute has seen
e2 + e2' e2 " no necessary antithesis between economic profit
es + e3' ese" S (35) and social gain. In emphasizing the long-range or
ill,
i2 +t =
i3 + i3= is = e33 /,R3
professional view, as against the short-range or
business view, it has sought to reconcile social
~~~~~profit
f2', f3', fi", f2', f3") may be *onie with
From these the new components (fi', * the and
sice profitgain,
oftothe balance sheet, to rec-
reco
determined. Now regarding the reflected waves (fI', f2', f3') as new oncile service and gain, to reconcile scientific
incident waves (dropping the primes) approaching the tower, an progress with humanitarianism and to reconcile the
exactly similar set of equations to those above obtain, except highest stimulation of individual achievement with
that R3 is now to be replaced by the resistance of the tower footing the widest distribution of social benefits. Before
and the lightning stroke surge impedance in parallel, that is the its members it has upheld the ideal of progress
through sharing, rather than beating the game. In
i3 + 4'} 8 4"t = e3" (R +r Z)/RZ (36) the larger concerns of society, as well as in the
Thus it iS possible to calculate the effect of the transient impedance acitesndrliosfisthu nsofnivul
of the counterpoise as a problem in successive reflections. But the members, the Institute has kept the faith, not alone
work involved is so excessive that the approximation given in the in the letter that killeth, but even more in the spirit
text appears to be sufficiently accurate under the circumstances, that maketh alive.

1172 ELECTRICAL ENGINEERING