Presented at the 34th A nnual Convention of

the A merican Institute of Electrical Engineers,

Atlantic City, N. J., June 27, 1918.

Copyright 1918. By A. I. E. E.

THE OXIDE FILM LIGHTNING ARRESTER

BY CHARLES P. STEINMETZ

ABSTRACT OF PAPER
A short history of lightning protection of electric systems is
given, as relating to the three successive types of electric circuits; the communication circuits, the power circuits of negligible
electrostatic capacity, and the high-power circuits containing
distributed capacity and inductance and capable of electric oscillation, leading to the three problems of discharging over-voltage
to ground, opening the power current which follows the discharge
and discharging so that no power current follows even for a
fraction of a half wave. It is shown that these problems are
solved by the spark gap to ground, by the use of non-arcing
metals in the multigap arrester, which opens the circuit at the
end of a half wave of current, and by the so-called "counter
e.m.f." type of arrester, represented by the aluminum cell and the
oxide film arrester.
It is shown that the necessity of taking care of recurrent discharges in high-power systems had led to the universal adoption
of the aluminum cell arrester in such systems, in spite of its disadvantage of requiring daily attendance in charging, and of
containing an electrolyte and oil.
In the oxide film arrester a type of arrester is presented which
has the same characteristics and therefore the same advantages
as the aluminum cell arrester, but does not require daily attendance and contains no liquids.
Its method of operation is explained, and its difference from
the aluminum cell arrester; the dielectric film, which punctures
under the discharge, and reseals after the discharge, is formed
from the solid materials between the terminal plates, compressed
PbO2, and therefore no spontaneous chemical action occurs
which dissolves the film, as in the aluminum cell, in which the
film forms from the aluminum electrode, gradually dissolves,
and therefore requires 'daily charging.
A short description of the construction of the oxide film arrester
is given, a record of its operation in industrial service for over
three years, and oscillograms showing the performance of this
arrester under recurrents, oscillations and under high-power impulses.

THREE periods can be distinguished in the development of

lightning arresters, corresponding to the three periods of
the use of electricity:
(1) Electric circuits of negligible power, telegraph and telephone.
(2) Electricpowercircuits of negligibleelectrostatic capacity;
871

872

STEINMETZ: OXIDE FILM ARRESTER

[June 27

d-c. lighting and railway, a-c. secondary and 2300-volt primary

distribution.
(3) High-voltage electric power circuits, transmission lines, etc.
(1) In electric circuits of negligible power, such as telegraph
and telephone circuits, a simple minute spark gap to ground
afforded protection by discharging lightning to ground, and was
sufficient until the recent years, when with the general introduction of electric power circuits the problem arose, in electric
communication circuits to take care of crosses with power circuits.
(2) In electric power circuits, a simple spark gap to ground
became insufficient for protection, since the power current, following the lightning discharge as arc, short-circuited the system
and burned up the arrester. The problem then arose, to safely
open the short circuit of the machine current to ground, through
the lightning arrester, after the lightning discharge has passed,
and to leave the arrester in operative condition to receive following lightning discharges.
Of the various devices developed heretofore, the magnetic
blow-out lightning arrester still is used in direct-current railway
circuits.
The first scientific investigation of this problem, is recorded
in the paper' by A. J. Wurts, presented before the A. I. E. E. at
the meeting of May 1894. Since that time all lightning arresters
for alternating-current power circuits of negligible electrostatic
capacity are based on the multigap principle between non-arcing
metals, whatever constructive forms the arrester may assumeas the present compression chamber lightning arrester. The
multigap arrester operates on the principle, that the lightning
discharge over the multigaps closes the circuit to ground, but
the power arc following the discharge extinguishes at the end
of the half wave of the alternating current, as the non-arcing
character of the gaps does not permit the reverse current of the
next half wave to start. The multigap arrester thus shortcircuits for a part of a half wave. It obviously is suited only for
alternating currents.
For years difficulties were met with the question of resistance;
without series resistance, in large systems the short-circuit power
1. TRANS. A. I. E. E., 1894, Vol. XI, p. 337, "Discriminating
Lightning Arresters and Recent Progress in Means for Protection Against
Lightning", by A. J. Wurts.

19181

STEINMETZ: OXIDE FILM ARRESTER

873

even of a part of a half wave may be sufficient to disable

and destroy the arrester, while the use of a series resistance,
while limiting the power current and thereby protecting the
arrester, also limited the discharge capacity and thereby reduced the protection. This problem was solved by the use of
multigaps shunting the series resistance, so that moderate discharges passed over the resistance, while high power lightning
discharges found a path without series resistance, over the
shunted gaps, and at the same time, the shunting resistance
made the power arc at the shunted gaps unstable and thus assisted in the extinction of the short circuit at the end of the
half wave.2
(3) As soon however, as circuits came into use, which had
considerable electrostatic capacity, such as high-voltage transmission circuits, extended underground cable systems, or lower
voltage circuits (including generator circuits) inductively connected with such circuits, the multigap arrester failed by frequently, or even usually destroying itself by the discharge, burning up.
In such circuits, oscillations between capacity and inductance
may occur, started by a lightning discharge or any internal disturbance such as switching etc., resulting in recurrent high-frequency oscillations, of which the arcing ground on a transmission
line is typical and probably best known. With such continual
discharges, often several per half wave, the multigap arrester
short-circuits at the first oscillation, for the remainder of the half
wave, and while the multigap functions properly and opens the
short circuit at the end of the half wave, the oscillation of the
next half wave again short-circuits, and so on, so that the effect
is that of a continuous short circuit, and no lightning arrester,
no matter how large, can dissipate the short-circuit power of
a big system for any appreciable time.
For such systems, in which recurrent high frequency oscillations, as arcing grounds, may occur, a lightning arrester is
necessary, which does not short-circuit the machine current
even for a fraction of a half wave, but merely discharges the over
voltage, the oscillation which, however high in voltage it may
be, is small in energy compared with the short-circuit power of
2. See A. I. E. E. TRANS., 1907, Vol. XXVI, p. 425," Protection Against
Lightning, and the Multigap Lightning Arrester" by D. B. Rushmore and
D. Dubois.

874

STEINMETZ: OXIDE FILM ARRESTER

[June 27

the system, as it represents only the stored energy of capacity

and inductance. The only arrester of this character heretofore
was the electrolytic, or aluminum cell lightning arrester, developed by E. E. F. Creighton, J. L. R. Hayden, F. W. Peek and
others. It acts towards an over-voltage discharge like a counter
e. m. f. equal to the normal circuit voltage, and the discharge
current passing through the arrester thus is the short-circuit
current of the over voltage, while the normal machine voltage
does not discharge, is held back and not disturbed. The aluminum cell arrester thus can discharge continual disturbances,
over-voltage oscillations occurring at every half wave, for a considerable time, half an hour to several hours, before it is endangered by the temperature rise due to the accumulated energy of
these discharges.
The aluminum cell arrester comprises a series of cells-usually
conical and stacked into each other-of aluminum electrodes
with an. electrolyte, of which neither the salt nor its ions appreciably dissolve alumina. In "forming" the cell, by an alternating current passing through it, the electrodes are coated by
a thin non-conducting film of alumina, which grows in thickness
until it holds back the impressed voltage. Any over-voltage
punctures this film, but the current passing through the puncture holes, again forms alumina andcloses the holes. Thus the
aluminum cell acts like a self repairing electrostatic condenser
of a disruptive strength equal to the impressed voltage: about
250 to 300 volts per cell.
The practical experience of the last ten years has proven the
aluminum cell arrester as the only type capable of affording protection in modern high-power circuits, and proven this so conclusively, as to lead to its universal adoption in such circuits
in spite of the inconveniences incident to the need of daily attention in charging, the use of a liquid electrolyte, and the difficulty
of testing the arrester without taking it apart, except by watching the appearance of the charging arc, or measuring the charging
current.
These inconveniences incident to the aluminum cell arrester
were well realized however, and as soon as the minor troubles
met with the aluminum cell arrester in the early years had been
overcome engineers went energetically to work on the problem
of developing a lightning arrester of the characteristics of the
aluminum cell arrester, but which does not require any attention

19181

STEINMETZ: OXIDE FILM ARRESTER

875

beyond that given to every apparatus in a well managed system,

that is, an occasional inspection, at least once or twice a year.

Numerous researches were made by the engineers whose splendid work I here acknowledge: Messrs. H. D. Brown, E. E. F.
Creighton, Crosby Field, V. E. Goodwin, J. L. R. Hayden, N. A.
Lougee, and G. B. Phillips. Some of these researches have not
yet led to results ready for communication, but I am glad to
present to you here as the result of the work of these men, and
more particularly of E. E. F. Creighton, Crosby Field and N.
A. Lougee, in the Oxide Film Lightning Arrester a new type of
lightning arrester, which has all the characteristics and advantages of the aluminum cell arrester, but does not require any
charging and thus requires no special attention, contains no
liquid electrolyte, no inflammable material, and like the aluminum cell arrester, can be located outdoors as well as indoors.
The oxide film arrester, like the aluminum arrester, acts like
a counter e. m. f. equal to the normal circuit voltage, freely discharging any over voltage, but holds back the normal machine
voltage. Thus the discharge is limited to the energy of the
over voltage, as in the aluminum arrester, and like the latter,
the oxide film arrester can continuously discharge recurrent
surges, such as arcing grounds etc., without endangering itself,
for a considerable time, sufficiently long to notice -and eliminate
the disturbance.
Compared with the almost entire absence of knowledge of lightning phenomena in electric circuits, under which Mr. Wurts had
to work in developing the non-arcing metal multigap lightning
arrester, our present knowledge of lightning phenomena is very
great. Nevertheless, there are so many disturbances in large
electric systems, which we cannot or only incompletely reproduce in our laboratories, that the final decision on the success,
that is, the effectiveness and permanence of a lightning arrester,
still is best given by the experience in industrial systems.
Therefore, after extensive laboratory tests had been completed
and had proven the oxide film arrester as of the same characteristic as the aluminum arrester, but requiring no special attention,
a number of industrial installations were made, and more added
the next year and the third year. Now, however, when a considerable number of installations of these arresters, for voltages
from 110 to 33,000, have been in successful operation, some for
over three years, and have proven their protective value and

876

STEINMETZ: OXIDE FILM ARRESTER

[June 27

their permanence, I consider it desirable to bring the arrester to

your attention.
Of the numerous tests made on the performance of the arrester,
it may be sufficient here to give only two, by the oscillograms
Figs. 1 to 4, showing the action on a recurrent oscillation, in Figs.
1 and 2, and on a single high-power impulse, Figs. 3 and 4.
The tests, oscillograms Figs. 1 and 2, were made in the usual
manner: a surge or continual oscillation was produced by a large
condenser, connected to an alternating-current supply and discharging over a spark gap through an inductance. The latter
was chosen so as to give a frequency of 1200 cycles to the oscillation, and thereby bring it well within the range of the oscillograph. This surge was impressed upon the apparatus to be
protected, a transformer energized by another alternating-current circuit. Fig. 1 shows the oscillogram without protection
of the transformer, and Fig. 2 the oscillogram with an oxide film
cell shunting the transformer and thereby protecting it.
In Fig. 1, the lowest curve shows the voltage of the 320-volt
47-cycle power supply circuit of the transformer, with the oscillations superimposed on it, rising to surge peaks of 2800 volts.
The upper curve shows the oscillating currents passing through
the transformer, rising to current peaks of 40 amperes. The
middle curve is absent, as no arrester is used in this test. In Fig.
2 however, where an oxide film cell is shunted across the transformer, the middle oscillogram shows the current oscillations
passing through the arrester, with peaks of 35 to 41 amperes.
The lower curve in Fig. 2 then shows again the circuit voltage
wave impressed upon the transformer, with the oscillation's cut
down by the oxide film cell. This voltage wave, the lower curve
in oscillogram Fig. 2, well illustrates the characteristic action of
this type of "counter-e. m. f. arrester"-to which the oxide film
arrester and the aluminum cell belong. The oscillation peaks
are sharply cut off at a maximum voltage of 60 per cent above
circuit voltage: the value for which the spark gap was set. As
the result, the oscillations are very greatly cut down from the
high values which they have in the unprotected circuit Fig. 1
(2800 volts), and become unsymmetrical. The half waves of oscillation in the same direction as the circuit voltage are greatly
reduced, by the limitation of the voltage to 60 per cent abovenormal, while the reverse half waves-which lower the instantaneous circuit voltage- are less affected. Corresponding thereto the discharge current through the arrester, the middle oscil-

PLATE XXXIV.
A. I. E. E.
VOL. XXXVII, 1918

_~~~~~-I
FI7G. 1-SURGE DISCHARGE AS IN FIG. 2-.No ARRIESTEwR IN CIRCuIr
Top vibrator-currcnt thrr... gh tra,,sfo,r,ver.
Bottoi

il,Tato r

ircuit

Itage.

FIG.2- STIRGE DISCHARGEA THROUGH ARRESTER CELL

Top vil,rator--60-cNvcle timiing wave.
Middle vibrator-currcnt through arrester.

Bottom vibrator--voltage acr.ss arrester.

STEINMETZ]

PLATE XXXV.
A. 1. E. E.
VOL. XXXVII, 1918

FIG. 3-SINGLE miMPuiLsE DiSCHARLGE-No ARRESTER IN CIRCUIT

Top vilbrator-60-cycle timning wave
Bottom vibrator-circo,it voltage.

[STE1NMb,TZI

FIG. 4-SiNCi,LE ImpuiLsE DISCHARGE AS IN F"IG. 3-ARRESTFR IN CIRCUIT

Top vibrator- 6-,v,51-cc timi lg wave.

Nliddile vibrat,r-curr,t through

Bott,n viheat-'

rl

g a,r,ss

arrester.

arre~ster.

PLATE XXXVI.
A. I. E. E.
VOL. XXXVII, 1918

FiG. 5-TYPE "0F" LiGHTNaING

ARRESTER CELL

[STEINMETZ]

FIG. 6-TYPE "0 F" LIGHTNING ARRESTERS WITH COVERs REMOVED

RATINGs 325-650, 900-1350, 2100-2600-VOLT ALTERNATING CURRENT

RESPECTIVELY

OR DIRECT CURRENT

8-FTYPE

FIG.
"O F" LGHTNING ARRESTER FOR OUTDOOR SERVICE INSTALLED ON A 33,000-VOLT CIRCUIT

[STEINMETZ]

ARRESTER

WITH

INTERIOR ARRANGEMENT

LIGHTNING

SIDE

REMOVED

OUTDOOR TYPE "OF"

FIG. 7-PHASE SECTION OF 15,000-25,000-VOLT
TO SHOW
OF HOUSING

(0

mx

1918]

STEINMETZ: OXIDE FILM ARRESTER

877

logram in Fig. 2, is unsymmetrical also: in the first and second

oscillation of Fig. 2, the first and third half wave of oscillating
voltage is cut off, and the first and third half wave of the oscillating discharge current therefore higher than the second half
wave of the oscillating discharge, which latter corresponds to a
half wave of oscillating voltage in opposition to the circuit voltage, therefore not raising the circuit voltage. The third oscillation of Fig. 2 happens to start with a half wave of oscillating
voltage in opposition to the circuit voltage, and the first halfwave oscillating discharge current through the arrester, in the
middle curve, thus is smaller than the second half wave; the
second half wave of the oscillation is cut down in the voltage
and therefore gives the maximum discharge current, 35 amperes
in this case.
This feature is well brought out by the oscillograms Figs. 1
and 2, due to the use of a different frequency, 60 cycles, for the
power supplying the oscillator. This caused the successive oscillations to occur at different parts of the half waves of the 47cycles circuit which was to be protected.
Fig. 3 and 4 then show the protective action of the oxide film
arrester on a 550-volt direct-current circuit, against a single
(non-oscillatory) impulse produced by opening a highly inductive
circuit (railway motor). Fig. 3 shows on the lower curve the
oscillogram of the impulse in the 550-volt circuit, rising to 7000
volts. The upper curve merely is a 60-cycle timing wave, to
enable measuring the duration of the impulse. Fig. 4 shows
the same circuit, with an arrester shunting it. The impulse
voltage rises to the value for which the discharge gap of the arrester is set, in this case 2200 volts. Then the arrester discharges,
and the voltage instantly drops back to normal, while a slowly
decreasing discharge current through the arrester dissipates
the magnetic energy of the impulse.
The cell of the oxide film arrester, shown in Fig. 5, consists of
two circular metal plates as electrodes, which are kept apart by
a porcelain ring, as shown in the figure. The space between the
electrodes inside of the porcelain ring, is filled with the active
material, lead peroxide PbO2, which is put in under moderate
pressure. This active material is a good conductor, but has the
characteristic, that by the action of an electric discharge, it is
converted in the path of the discharge, into a lower oxide, which
is an insulator. Thus when an alternating current is passed
through such a cell, the active material at the electrodes grad-

878

STEINMETZ: OXIDE FILM ARRESTER

[June 27

ually converts into a non-conductor, and forms a thin insulating

film at the electrode. This grows in thickness, until it cuts off
the further flow of current and holds back the voltage, about
250 to 300 volts per cell. Then only a small leakage current, of
a few milliamperes, passes at normal voltage, but if an over voltage of any kind appears at the cell, the insulating film of lead
oxide punctures and freely discharges through the lead peroxide,
but in doing so, converts the surface of the lead peroxide in
the path of the discharge into the lower non-conducting oxide,
and thereby closes the puncture holes, repairs or reseals the
film.
In manufacture, naturally, just as in the aluminum cell arrester,
the insulating film is not produced after assembly of the cell by
the slow process of passing a current through, but the film is put
on before assembly, in the oxide film arrester, by dipping the
plates in a suitable insulating varnish, which gives them a coating
just thick enough to hold back the circuit voltage. Then after
assembly, voltage is put on the cell for testing it and sealing any
holes or defects which may exist in the varnish film.
In the oxide film arrester, the electrodes have nothing to do
with the arrester action, and any suitable material can be used.
First we used brass, but now use sherardized iron, the latter
having a higher melting point and thus standing high power discharges, which would melt holes in the brass electrodes.
In this arrester, the action, which holds back the normal voltage but passes freely an over voltage, thus resides in the active
material between the electrodes, and it is this material which
forms and reforms the film. As this material is a solid, no chemical action occurs such as the gradual dissolution of the alumina
film in the aluminum cell arrester, but the film remains intact
permanently, and thus no daily "charging", that is, repairing
of the film, is required.
A number of such cells, depending on the voltage of the circuit,
are piled on top of each other, with a spark gap in series, and, for
low and moderate voltages, incased as shown in Fig. 6.
As the cells are hermetically sealed, by the metal of the electrodes being spun over the porcelain separating ring, the cells
can be installed outdoors as well as indoors, requiring in outdoor
installation merely some protection by petticoats, as shown in
Fig. 7, to keep the rain from short-circuiting the cells. Fig. 8
shows such an outdoor installation of a 33-kv. arrester, with threephase stacks and the ground stack of cells, protected against the

1918]

STEINMETZ: OXIDE FILM ARRESTER

879

weather by metal petticoats. Fig. 8 shows also the spark gaps on

the line side of the arresters. They are protected sphere gaps, to
give instantaneous discharge, with a horn attachment to allow
the arc to flare up and thereby help in its extinction.
As well known, the plain horn gap has the disadvantage of
requiring an appreciable -though a very short (microseconds)
-time for discharge, and an extremely sudden high voltage, as a
very steep wave front, thus may pass the horn gap and flash over
elsewhere. Therefore in modern high-voltage lightning arresters
the horn gap is shunted by a properly proportioned sphere gap,
the latter being "instantaneous" in its action. In outdoor use
however, rain lowers the discharge voltage of the sphere gap,
and thus requires a setting which gives a higher discharge voltage in dry weather than necessary. Therefore a protected sphere
gap has been designed, which overcomes this disadvantage in
the open sphere gap, and is shown in Fig. 8.
The need of using this spark gap in series with the arrester,
is the only still remaining undesirable feature which the oxide
film arrester shares with the aluminum cell -arrester, the multigap arrester and other types. While by the work of Mr. F. W.
Peek, on the time lag of electric discharges3, the means have been
given to make the discharge gap "instantaneous", that is, faster
than any other discharge path over gaps or through insulation
in the system, so that the arrester takes care effectively of any
over voltage above its discharge voltage, it does not discharge
voltages lower than the discharge voltage of its spark gap, even
if these lower voltages may involve some danger to the system by
their high frequency. Such low voltages, while they cannot
endanger the main insulation between circuit and ground, may,
if of sufficiently high frequency, lead to local accumulations of
voltage across inductive parts of the circuit, as regulators, current
transformers, end turns and coils of generators and transformers, and there cause damage by puncturing insulation between
turns and causing internal short circuits.
Against these high frequency disturbances of moderate voltage, the only existing protection is the addition to the arrester
of a capacity discharge path permanently connected from the
circuit to ground. Such capacity path should be without resistance to flatten steep wave fronts, and contain a moderate series
3. A. I. E. E. TRANS. 1915, Vol. XXXIV, Part II, p. 1857, The Effect of Transient Voltages on Dielectrics by F. W. Peek, Jr.

880

STEINMETZ: OXIDE FILM ARRESTER

[June 27

resistance, to dissipate high-frequency energy and stop cumulative oscillations in their beginning. Before I leave the field of
electrical engineering, I hope still to see an arrester, of the type
of the oxide film or the aluminum cell, which has no spark gap,
but is permanently shunted across the circuit, and thus capable
of taking care not only of over voltages, but equally well of steep
wave fronts and high-frequency oscillations, even if of lower than
the circuit voltage. Such an arrester then would give universal
protection.