EE4270

High Voltage Breakdown and Testing
Outline Syllabus
Level 4, Semester 2
Professor J R Lucas
May 2010
1. High Voltage Breakdown Phenomena [6 hrs]
Breakdown Characteristic in gases:
Electron Avalanche Mechanism, Townsend Breakdown Process,
Streamer Mechanism, Time lags of Spark breakdown.
Corona Discharges, Mechanism of corona formation, Power Loss
due to Corona
Breakdown in Liquids:
Breakdown of Commercial liquids; Breakdown due to gaseous
inclusions, liquid globules, solid particles;
Breakdown of Solid Insulating Materials:
Electro-mechanical breakdown, Breakdown due to internal
discharges, Surface Breakdown, Thermal Breakdown, Electro-
chemical Breakdown, Chemical Deterioration, Breakdown of
Composite Insulation.
2. High Voltage Cables [6 hrs]
Power loss in the cable: Dielectric loss, Conductor loss, Sheath loss,
Intersheath Loss, Cross-bonding of Cables
Impregnated paper insulation: Properties required, Principle
underlying the design, Paper insulated power cables,
Single core and three core cables: Insulation Resistance,
Capacitance, Copper Space Factor
Dielectric stress in a single core cable: Cable Grading for Uniform
Stress Distribution, Capacitance Grading, Intersheath Grading
Pressurised high voltage cables: Oil-pressure cables, Gas-pressure
cables, External Pressure Cables, Internal Pressure Cables
Thermal design of cables: Current rating, Thermal Resistance -
single-core cables, three-core cables, protective coverings, ground
around cable, Cables exposed to air
High voltage bushings: Simple cylindrical bushing, Condenser
bushing
3. High Voltage Generators for Testing [6 hrs]
Generation of High Alternating Voltages:
Cascade arrangement of transformers
Resonant Transformers
High frequency high voltages
Generation of High Direct Voltages:
Rectifier circuits
Voltage Multiplier Circuits
Electrostatic generators: Van de Graeff generator, Sames
Generator
4. High Voltage Measurements and Testing [6 hrs]

Electrostatic voltmeter, sphere gaps, potential dividers,
matching, peak reading meters, Klydonograph
Type tests, Sample Tests, Routine Tests
Oscilloscopes for the measurement of fast transients
Measurements of capacitance and loss tangent: High Voltage
Schering Bridge, Dielectric loss measurement,
Detection of internal discharges
Measurement of dielectric constant and dissipation factor of a
liquid dielectric
General tests carried out on High voltage equipment.
Testing of solid dielectric materials.
1
Breakdown of
Gaseous Insulation
3 J R Lucas Breakdown of Gaseous Insulation
Electrical Insulating Materials
referred to as Dielectrics
electrostatic fields can remain almost indefinitely
offer very high resistance to the passage of direct currents
cannot withstand an infinitely high voltage.
○ when applied voltage across dielectric exceeds a critical value the
insulation will be damaged
may be gaseous, liquid or solid in form.
1 J R Lucas Breakdown of Gaseous Insulation
Can a person swim one length without hitting some one ?
If the pool is almost empty
If the pool is crowded
Will he be able get sufficient momentum in these cases to hurt someone?
1 J R Lucas Breakdown of Gaseous Insulation
1.1 Ionisation of Gases
Gaseous dielectrics are not free of electrically charged particles, including
free electrons.
Free electrons
may be caused by irradiation or field emission
can lead to a breakdown process to be initiated.
on the application of an electric field are accelerated from cathode to
anode by the electric stress applying a force on them.
Force = mass× acceleration, Force = charge× electric field
acquire a kinetic energy (½ mu
2
) as they move through the field.
moving towards the anode, collide with gas molecules present between
the electrodes.
part of the kinetic energy of the electrons is lost in these collisions, and
part is transmitted to the neutral molecule.
3 J R Lucas Breakdown of Gaseous Insulation
Energy is usually expressed as a voltage, E
i
= e V
i

in electron-volt (eV) as the energies involved are extremely small,
where e is the charge on an electron = 1.6 x 10
-19
C.
1 e V = 1.6 x 10
-19
J
If molecule gains sufficient energy (more than
the ionisation energy E
i
), it may ionise by
collision.
Mean number of ionising collisions by one
electron per unit drift across the gap is not a
constant but subject to statistical fluctuations.
The newly liberated electron and the impinging
electron are then accelerated in the field and an
electron avalanche is set up.
3 J R Lucas Breakdown of Gaseous Insulation
Further increase in voltage results in additional ionising processes.
Ionisation increases rapidly with voltage once these secondary processes
take place, until ultimately breakdown occurs.
It is worth noting that in uniform fields, the ionisation present at voltages
below breakdown is normally too small to affect engineering applications.
In non-uniform fields, however,
considerable ionisation may be present in
the region of high stress, at voltages well
below breakdown, constituting the well
known corona discharge.
1 J R Lucas Breakdown of Gaseous Insulation
1.1.1 Ionisation processes in gas discharges
Electrical breakdown of a gas is caused by various processes of ionisation.
gas processes involving the collision of electrons, ions and photons
with gas molecules, and
electrode processes which take place at or near the electrode surface
[Electrons can be emitted from the cathode at stresses around 100 –
1000 kV/cm due to field emission].
Ionisation is the process by which an electron is removed from an atom,
leaving the atom with a nett positive charge (positive ion).
Since an electron in the outermost orbit is subject to the least attractive
force from the nucleus, it is the easiest removed by any of the collision
processes.
1 J R Lucas Breakdown of Gaseous Insulation
The energy required to remove an outer
electron completely from its normal state in
the atom to a distance well beyond the
nucleus is called the first ionisation
potential.
The reciprocal process of an electron falling
from a great distance to the lowest
unoccupied orbit is also possible. In this
case, a photon will be emitted having the
same energy as previously absorbed.
3 J R Lucas Breakdown of Gaseous Insulation
1.1.2 Relevant gas ionisation processes
(i) Ionisation by simple collision
When the kinetic energy of an electron (½ mu²), in collision with a neutral
gas molecule exceeds the ionisation energy (E
i
= e V
i
) of the molecule,
then ionisation can occur.
(i.e. the necessary but not necessarily sufficient condition is ½ mu² > E
i
)
M + e
-
(½ mu²) M
+
+ 2 e
-
In general, a positive ion and 2 slow moving electrons will result.
The probability of this process is zero for electron energies equal to the
ionisation energy E
i
, but increases almost linearly at first, and then
gradually with electron energy up to a maximum.
When the gas molecules are bombarded with electrons, other electrons
bound to atoms may be freed by the collision with the high energy
electron.
3 J R Lucas
001
0
0
2
0
0
3
0
0
4
0
0
5
0
0
6
0
0
e
l
e
c
t
r
o
n

e
n
e
r
g
y

(
e
V
)
i
o
n
i
s
a
t
i
o
n

p
r
o
b
a
b
i
l
i
t
y
Breakdown of Gaseous Insulation
Figure 1.1 - Ionisation probability curve in air
4 J R Lucas Breakdown of Gaseous Insulation
Ratio of, the electrons given by collision to the primary electrons, depend
mainly on the energy of the primaries.
This is maximum at primary electron energies of about 200 – 500 eV.
For lower energy values, the energy transferred may not be sufficient to
cause electrons to escape from the surface of the molecules, and thus the
probability of ionisation is small.
For much higher values of primary energies, the energy of the impinging
electron would be sufficient for this electron to penetrate the surface
deeper into the molecule, so that again the chance of escape of other
electrons decreases.
(ii) Excitation
In the case of simple collision, the neutral gas molecule does not always
gets ionised on electron impact.
The molecule may be left in an excited state M
*
, with energy E
e
.
2 J R Lucas Breakdown of Gaseous Insulation
M + e
-
(½ mu²) → M
*
+ e
-
This excited molecule can subsequently give out a photon of frequency v
with energy emitted h v.
Energy is given out when the electron jumps from one orbit to the next.
M
*
 M + h v
where h = Planck's constant = 6.624 x 10
-34
J s
(iii) Ionisation by Double electron impact
If a gas molecule is already raised to an excited state (energy E
e
) by a
previous collision, then ionisation of can occur by a collision with a
relatively slow electron.
3 J R Lucas Breakdown of Gaseous Insulation
Less energy would be required than the ionisation energy, but the energy
must exceed the additional energy required to attain the ionisation energy.
(i.e. necessary but not sufficient condition is ½ mu² > E
i
- E
e
)
M
*
+ e
-
(½ mu²) → M
+
+ 2 e
-
(iv) Photo-ionisation
A molecule in the ground state can be ionised by a photon of frequency 
provided that the quantum of energy emitted h ν (by an electron jumping
from one orbit to another), is greater than ionisation energy of molecule.
(i.e. h ν > E
i
, where h = Plank's constant =
6.624 x 10
-34
joule)
M + h  → M
+
+ e
-
(v) Electron Attachment
If a gas molecule has unoccupied energy levels in
its outermost group, then a colliding electron may
2 J R Lucas Breakdown of Gaseous Insulation
take up one of these levels, converting the molecule into a negative ion
M
-
.
M + e
-
→ M
-
The negative ion thus formed would be in an excited state, caused by the
excess energy.
Note: Electron attachment decreases the number of free electrons, unlike
ionisation which increases the free electrons.
(vi) Electron detachment
This occurs when a negative ion gives up its extra electron, and becomes a
neutral molecule.
M
-
→ M + e
-
(vii) Other Processes
The above discussed processes are the most important in relation to the
gas discharge phenomena.
1 J R Lucas Breakdown of Gaseous Insulation
Other possible gas processes include ion-atom collisions, excited atom-
molecule collisions, and atom-atom collisions.
It should be noted that collisions between ions and atoms rarely result in
ionisation, due to the relatively slow interaction time, which allows the
internal motion of the atomic system to adjust itself gradually to the
changing condition without any energy transition occurring.
In order to cause ionisation of a neutral unexcited atom of its own kind, a
positive ion must possess energy of at least 2 eV.
Normally ions and atoms having such energies are encountered only in
high current arcs and thermonuclear discharges.
3 J R Lucas Breakdown of Gaseous Insulation
1.2Breakdown Characteristic in gases
Two mechanisms of breakdown in gasses are known which explain the
behaviour under different conditions.
These are the Townsend and streamer mechanisms.
Both depend on an initial electron avalanche mechanism.
1.2.1 Electron Avalanche Mechanism
Suppose a free electron exists (caused by some external effect such as
radio–activity or cosmic radiation) in a gas where an electric field exists.
If the field strength is sufficiently high, then it is likely to ionize a gas
molecule by simple collision resulting in 2 free electrons and a positive
ion.
These 2 electrons will be able to cause further ionization by collision
leading in general to 4 electrons and 3 positive ions.
3 J R Lucas Breakdown of Gaseous Insulation

The process is cumulative, and the number of free electrons will go on
increasing as they continue to move under the action of the electric field.
The swarm of electrons and positive ions produced in this way is called an
electron avalanche.
A
n
o
d
e

+
C
a
t
h
o
d
e



½mu
2
Anode + C
a
t
h
o
d
e



Anode + C
a
t
h
o
d
e



Anode + Cathode –
F
i
g
u
r
e

1
.
2



A
v
a
l
a
n
c
h
e

P
r
o
c
e
s
s

b
u
i
l
d

u
p

A
n
o
d
e

+
C
a
t
h
o
d
e



Molecule
Electron
Positive ion
1 J R Lucas Breakdown of Gaseous Insulation
In the space of a few millimetres, it may grow until it contains many
millions of electrons.
F
i
g
u
r
e

1
.
2



A
v
a
l
a
n
c
h
e

P
r
o
c
e
s
s

b
u
i
l
d

u
p

3 J R Lucas Breakdown of Gaseous Insulation
1.2.1.1 Mathematical Analysis [Townsend Breakdown Process ]
When the voltage applied across a pair of electrodes is increased, the non-
self-sustaining current throughout the gap increases slowly.
The electrons emitted from the cathode move through the gas with an
average velocity determined by their mobility for the field strength
existing for the particular value of voltage.
Impact ionization by electrons is probably the most important process in
the breakdown of gasses, but this process alone is not.
Let n
0
= No. of electrons/second emitted from the cathode,
n
x
= No. of electrons/second moving at a distance x from cathode
[n
x
> n
0
due to ionising collisions in gap]
α = number of ionising collisions, on average, made by one
electron per unit drift in the direction of the field.
[Townsend's first ionisation coefficient]
3 J R Lucas Breakdown of Gaseous Insulation
then 1/α=average distance traversed in the field direction between
ionising collisions.
Consider a laminar of thickness dx at a distance x from the cathode. The
n
x
electrons entering the laminar will traverse it in the presence of the
applied field E.
The ionising collisions generated in the gas gap will be proportional to
both dx and to n
x
.
Anode + C
a
t
h
o
d
e



x

d
x

5 J R Lucas Breakdown of Gaseous Insulation
Thus dn
x
∝ n
x
∝ dx
Therefore dn
x
= α . n
x
. dx (from definition of α)
1 J R Lucas Breakdown of Gaseous Insulation
Rearranging and integrating gives
e
.
n
=
n

x . = )
n
/
n
(
dx =
n
n
d

x
0 x
0 x
e
x
0
x
x
n


n
x
0
α
α
α
log
∫ ∫
If the anode is at a distance x = d from the cathode, then the number of
electrons n
d
striking the anode per second is given by
n
d
= n
0
. e
αd
Therefore, on the average, each electron leaving the cathode produces (n
d
- n
0
)/n
0
new electrons (and corresponding positive ions) in the gap.
In the steady state, the number of positive ions arriving at the
cathode/second must be exactly equal to the number of newly formed
electrons arriving at the anode.
1 J R Lucas Breakdown of Gaseous Insulation
Thus the circuit current will be given by
I = I
0
. e
αd
where I
0
is the initial photo-electric current at the cathode.
In the actual breakdown process, the electron impact ionization is attended
by secondary processes on the cathode, which replenish the gas gap with
free electrons, with every newly formed avalanche surpassing the
preceding one in the number of electrons.
Consider the current growth equations with secondary mechanism also
present.
Let γ = No. of secondary electrons (on average) produced at the cathode
per ionising collision in the gap.
[Townsend's second ionisation coefficient]
n
0
=No. of primary photo-electrons/second emitted from the cathode
n
0
' =No. of secondary electrons/second produced at the cathode
3 J R Lucas Breakdown of Gaseous Insulation
n
0
" =Total number of electrons/second leaving the cathode
Then n
0
" = n
0
+ n
0
'
On the average, each electron leaving the cathode produces [e
αd
- 1]
collisions in the gap, giving the number of ionising collisions/second in
the gap as n
0
" (e
αd
- 1).
Thus by definition
) 1 -
e
( "
n
n
=
d
0
0
α
γ
'

γ
γ
α
α
. ) 1 -
e
( "
n
+
n
= "
n
that so
'
n
+
n
= "
n
but
) 1 -
e
( "
n
=
n
giving
d
0 0 0
0 0 0
d
0 0′

1 J R Lucas Breakdown of Gaseous Insulation
) 1 -
e
( - 1
n
= "
n

result the gives This
d
0
0
α
γ
Similar to the case of the primary process (with α only), we have
) 1 -
e
( - 1
e n
=
e
"
n
=
n

d
d
0 d
0 d
α
α
α
γ
Thus, in steady state, the circuit current I will be given by
) 1 -
e
( - 1
e I
= I
d
d
0
α
α
γ
This equation describes the growth of average current in the gap before
spark breakdown occurs.
As the applied voltage increases, e
αd
and γ e
αd
increase until γ e
αd
→ 1 ,
when the denominator of the circuit current expression becomes zero and
the current I → ∞ .
1 J R Lucas Breakdown of Gaseous Insulation
In this case, the current will, in practice, be limited only by the resistance
of the power supply and the conducting gas.
This condition may thus be defined as the breakdown and can be written
as
γ (e
αd
- 1) = 1
This condition is known as the Townsend criteria for spark breakdown.
The avalanche breakdown develops over relatively long periods of time,
typically over 1 μs and does not generally occur with impulse voltages.
3 J R Lucas Breakdown of Gaseous Insulation
1.2.1.2 Determination of Townsend's Coefficients α and γ
Townsend's coefficients are determined in an ionisation chamber which is
first evacuated to a very high vacuum of the order of 10
-4
and 10
-6
torr
before filling with the desired gas at a pressure of a few torr.
The applied direct voltage is about 2 to 10 kV, and the electrode system
consists of a plane high voltage electrode and a low voltage electrode
surrounded by a guard electrode to maintain a uniform field.
The low voltage electrode is earthed through an electrometer amplifier
capable of measuring currents in the range 0.01 pA to 10 nA.
The cathode is irradiated using an ultra-violet lamp from the outside to
produce the initiation electron.
The voltage current characteristics are then obtained for different gap
settings.
At low voltage the current growth is not steady.
1 J R Lucas Breakdown of Gaseous Insulation
Afterwards the steady Townsend process develops as shown in figure 1.3.
1 J R Lucas Breakdown of Gaseous Insulation
self sustaining
discharges
non-self
sustaining
discharges
Disch
arge
curre
nt I
V
o
V
1

V
2
voltage V
Figure 1.3 -
Typical Current
growth
brea
kdo
wn
1 J R Lucas Breakdown of Gaseous Insulation
From the Townsend mechanism, the discharge current is given by
1 » d when α
γ
α
α
α

e I
) 1 -
e
( - 1
e I
= I
d
0
d
d
0

This can be written in logarithmic form as
ln I = α d + ln I
0
y = m x + c
1 J R Lucas Breakdown of Gaseous Insulation
gap spacing
d
ln I Intercept =ln I
o
gradient = α α d << 1 Figure 1.4 - Variation of ln I
vs d
3 J R Lucas Breakdown of Gaseous Insulation
From a graph of ln I vs d, the constants α and I
0
can be determined from
the gradient and the intercept respectively, as in figure 1.4.
Once I
0
and α are known, γ can be determined from points on the upward
region of the curve. The experiment can be repeated at different pressures
if required.
1.2.1.3 Breakdown in electronegative gases
In the above analysis, electron attachment to
neutral molecules was not considered.
Electron attachment removes free electrons
and thus gives gases very high dielectric
strengths.
The gases in which electron attachment
occurs are electro-negative gases.
3 J R Lucas Breakdown of Gaseous Insulation
An attachment coefficient η can be defined, analogous with α, as the
number of attachments per electron per unit drift in the direction of the
field.
Under these conditions, the equation for the average current growth in a
uniform field can be shown to be as follows.
[ ] 1 1 -
e
-
-
-
-
e
-

I
= I
d ) - (
d ) - (
0
η α
η α
η α
α
γ
η α
η
η α
α
]
]
]

The corresponding criteria for spark breakdown is
[ ]
1 1 =
e
-

d ) - (

η α
η α
α
γ
1 J R Lucas Breakdown of Gaseous Insulation
1.2.2 Paschen's Law
When electrons and ions move through a gas in a uniform field E and gas
pressure p, their mean energies attain equilibrium values dependant on the
ratio E/p; or more precisely
α/p = f
1
(E/p) and γ = f
2
(E/p)
For a uniform field gap, the electric field E = V/d.
Thus applying Townsend's criterion for spark breakdown of gases gives
1 1 = ) -
e
(
d α
γ
which may be written in terms of the functions as
1 1 =
e
d p
V
f
d p
V
f d p
2
1
]
]
]

,
`

.
|

,
`

.
|
3 J R Lucas Breakdown of Gaseous Insulation
This equation shows that the breakdown voltage V is an implicit function
of the product of gas pressure p and the electrode separation d.
i.e. V = f (p.d)
In the above derivation the effect of temperature on the breakdown
voltage is not taken into account.
Using the gas equation
pressure × volume = mass R × absolute Temperature
pressure = density  R × absolute Temperature.
Thus it can be deduced that the correct statement of the above expression
is V = f(ρ.d), where ρ is the gas density.
This is the statement of Paschen's Law.
1 J R Lucas Breakdown of Gaseous Insulation
Under constant atmospheric conditions, it is experimentally found that the
breakdown voltage of a uniform field gap may be expressed in the form
V = A . d + B . d where d is the gap spacing
For air, under normal conditions,
A = 24.4 kV/cm and B = 6.29 kV/cm
1/2.
[The breakdown voltage gradient is about 30 kV/cm in uniform fields for
small gaps of the order of 1 cm, but reduces to about 25 kV/cm for large
gaps of several meters]
Figure 1.5 shows a typical breakdown vs spacing characteristic.
gap
spacing d
(cm)
breakd
own
voltage
(kV)
Figure 1.5 - Breakdown
characteristic
1 J R Lucas Breakdown of Gaseous Insulation
This variation with spacing can be modified using Paschen's Law to
include the variation with gas density as follows.
) d ( B + ) d ( A =
) d ( .
B
+ d .
A
= V
0
1/2
0
2
1
2
1
δ δ
ρ
ρ
ρ
ρ
where δ = relative density (or gas density correction factor).
This equation is true for gap spacings of more than 0.1 mm at N.T.P.
However, with very low products of pressure × spacing, a minimum
breakdown voltage occurs, known as the Paschen's minimum.
pressure x spacing
(cm)
breakdown
voltage (kV)
Figure 1.6 - Breakdown characteristic for low p.d
values
1 J R Lucas Breakdown of Gaseous Insulation
This can be explained in the following manner.
Consider a gap of fixed spacing, and let the pressure decrease from a point
to the right of the minimum (Figure 1.6).
The density therefore, decreases and an electron moving in the field
consequently makes fewer collisions with gas molecules as it travels
towards the anode.
Since each collision results in a loss of energy, it follows that a lower
electric stress suffices to impart to electrons the kinetic energy (½ m u
2
)
required to ionize by collision.
Near the minimum of the characteristic, the density is low and there are
relatively few collisions.
It is necessary now to take into account the fact that an electron does not
necessarily ionize a molecule on collision with it, even if the energy of the
electron exceeds the ionisation energy.
1 J R Lucas Breakdown of Gaseous Insulation
The electrons have a finite chance of ionizing, which depends upon its
energy.
If the density and hence the number of collisions is decreased, breakdown
can occur only if the chance of ionising is increased, and this accounts for
the increase in voltage to the left of the minimum.
It is worth noting that if the density is fixed, breakdown to the left of the
minimum occurs more readily (i.e. at lower voltage) at longer distances.
Typically the voltage minimum is 300 V and occurs at a product or p.d of
5 torr mm, or at a gap of about 0.06 mm at N.T.P.
At very low pressures, and at very high pressures (compared with
atmospheric), Paschen's Law fails.
Also, Paschen's Law is valid for temperatures below about 1100
0
C.
1 J R Lucas Breakdown of Gaseous Insulation
A further increase in temperature ultimately results in the failure of
Paschen's Law because of thermal ionisation.
It can be seen from the breakdown characteristic, that for a constant gap
spacing, the breakdown voltage, and hence the breakdown stress required,
is very much higher than at atmospheric conditions, under very high
pressure and under very low pressures (high vacuum) conditions.
Thus both high vacuum as well as high gas pressure can be used as
insulating media.
[The vacuum breakdown region is that in which the breakdown voltage
becomes independent of the gas pressure]
Under normal conditions, operation is well to the right of Paschen's
minimum.
Gas V
min
(V) p.d at V
min
(torr-cm)
2 J R Lucas Breakdown of Gaseous Insulation
Air 327 0.567
Argon 137 0.9
Hydrogen 273 1.15
Helium 156 4.0
Carbon dioxide 420 0.51
Nitrogen 251 0.67
Nitrous oxide 418 0.5
Oxygen 450 0.7
Sulphur dioxide 457 0.33
Hydrogen Sulphide 414 0.6
Table 1.1 - Minimum spark breakdown voltages
The table 1.1 gives the minimum sparking potential for various gases.


– –
– –
– – –
– – –
– – –
– – –

– – – –
– –
– – – –
– –
– –
– –



+
+
+ +
+ +
++ + − −
++ +
+ + + +
+ + +
+ + +
+ +
+ + +
+ +
+ + +
++
+ +
+ +
+ +
++
+++
+
+ +
++
+ +
+
++
+
++
+
++
+
+
+
+
+
+
+


catho
de
+
anod
e
Figure 1.7 -
Streamer
Mechanism
2 J R Lucas Breakdown of Gaseous Insulation
1.2.3 Streamer Mechanism
This type of breakdown mainly arises due to the added effect of the space–
charge field of an avalanche and photo–electric ionization in the gas volume.
While the Townsend mechanism predicts a very diffused form of discharge, in
actual practice many discharges are found to be filamentary and irregular.
The Streamer theory predicts the development of a spark discharge directly from
a single avalanche.
The space charge produced in the avalanche causes sufficient distortion of the
electric field that those free electrons move towards the avalanche head, and in
so doing generate further avalanches in a process that rapidly becomes
cumulative.


– –
– –
– – –
– – –
– – –
– – –

– – – –
– –
– – – –
– –
– –
– –



+
+
+ +
+ +
++ + − −
++ +
+ + + +
+ + +
+ + +
+ +
+ + +
+ +
+ + +
++
+ +
+ +
+ +
++
+++
+
+ +
++
+ +
+
++
+
++
+
++
+
+
+
+
+
+
+

+ + + + +
+ +++
+ + + +
+ ++
++ + + +
+++
++ + + +
+++
++ + + +
++
++ + + +
+
+ ++++
+ +++
+ ++
+
++ +
+
+++
++
+++
+
++
+
++
+
+


cathod
e
Figure 1.8 -
Streamer
Mechanism
+
anode
2 J R Lucas Breakdown of Gaseous Insulation
As the electrons advance rapidly, the positive ions are left behind in a relatively
slow–moving tail.The field will be enhanced in front of the head.
Just behind the head the field between the electrons and the positive ions is in
the opposite direction to the applied field and hence the resultant field strength
is less.
Again between the tail and the cathode the field is enhanced. (Figure 1.7)
Due to the enhanced field between the head and the anode, the space
charge increases, causing a further enhancement of the field around the
anode.
The process is very fast and the positive space charge extends to the
cathode very rapidly resulting in the formation of a streamer.
Figure 1.8 shows the streamer breakdown process.
1.2.4 Factors affecting the breakdown voltage a Vacuum gap
+ + + + +
+ +++
+ + + +
+ ++
++ + + +
+++
++ + + +
+++
++ + + +
++
++ + + +
+
+ ++++
+ +++
+ ++
+
++ +
+
+++
++
+++
+
++
+
++
+
+

2 J R Lucas Breakdown of Gaseous Insulation
Vacuum is ideally the best insulator, with breakdown strengths of the
order of 10
4
kV/c.
The breakdown voltage of a high vacuum is the voltage which when
increased by a small amount will cause the breakdown of the gap that was
held at that voltage for an infinite time.
However, this definition is not always practicable as the breakdown is
affected by many factors.
(i) Electrode Separation
For vacuum gaps less than about 1 mm, the breakdown voltage is
approximately proportional to the length, all other parameters remaining
constant.
This gives a constant breakdown strength.
1 J R Lucas Breakdown of Gaseous Insulation
For these small gaps, the breakdown stress is relatively high, being of the
order of 1 MV/cm.
Field emission of electrons probably plays an important part in the
breakdown process.
V = k . d for d < 1 mm
gap spacing
(cm)
1 2
3 4
breakdo
wn
75
0
50
0
25
0

0
actual
breakdown
voltage (kV)
Breakdown
stress
(kV/cm)
Figure 1.9 -
Breakdown
characteristic
breakdown voltage (kV)
if caused by field
emission only
3 J R Lucas Breakdown of Gaseous Insulation
For gaps greater than about 1 mm, the breakdown voltage does not
increase at an equal rate and the apparent breakdown stress for longer
gaps is much reduced, being about 10 kV/cm at a spacing of 10 cm.
[Apparent stress is defined as the voltage required to cause breakdown
divided by the distance between the electrodes] Figure 1.9.
Cranberg has shown theoretically that for longer gaps, it is the product of
breakdown voltage and breakdown stress that remains constant.
V . E = k
1
for d > 1 mm.
where the constant k
1
depends on the material and surface conditions of
the electrodes.
For an Uniform field gap, E = V/d, so that
V = k
2
d
1/2
Or in a more general form, both regions can be expressed by the equation
3 J R Lucas Breakdown of Gaseous Insulation
V = k d
x
where x = 0.5 for long gaps > 1 mm
= 1 for gaps < 1mm
(ii) Electrode Effects
Conditioning
The breakdown voltage of a gap increases on successive flashovers, until
a constant value is reached.
The electrodes are then said to be conditioned.
This increase in voltage is ascribed to the burning off by sparking of
microscopic irregularities or impurities which may exist on the electrodes.
When investigating the effect of various factors on breakdown, the
electrodes must first be conditioned in such a way that reproducible
results are obtained.
3 J R Lucas Breakdown of Gaseous Insulation

5 10 15
20 25
Breakdo
wn
Voltage
(kV)

40

30

20

10

0
Number
of
breakdo
wn
Figure 1.10 – Breakdown
characteristic
××××××××××××××××××××××××××××
1 J R Lucas Breakdown of Gaseous Insulation
The effect of conditioning is shown.
The breakdown voltage of conditioned electrodes or gaps is very much
reproducible than otherwise, and hence breakdown values that are
normally obtained experimentally are those of conditioned electrodes.
Unconditioned electrodes may have breakdown values a low as 50% of
the breakdown voltage with conditioned electrodes.
(iii) Material and Surface Finish
Electrode
Material
Voltage across
1 mm gap (kV)
Steel 122
Stainless Steel 120
Nickel 96
Monel metal 60
Aluminium 41
2 J R Lucas Breakdown of Gaseous Insulation
Copper 37
Table 1.2 - Effect of electrode material on breakdown
The electrode surfaces form the physical boundaries between which the
breakdown finally takes place.
Thus it is not surprising to find that the breakdown strength of a given size
of gap is strongly dependant on the material of the electrodes.
In general, the smoother the surface finish, the greater the breakdown
voltage.
(iv) Surface contamination
Presence of contamination in the test cell reduces the breakdown voltage
sometimes by as much as 50% of the clean electrode value.
(v) Area and configuration of electrodes
2 J R Lucas Breakdown of Gaseous Insulation
Increasing the area of the electrodes makes it more difficult to maintain a
given breakdown voltage. Thus breakdown voltage decreases slightly
with increase in surface area. For example, electrodes of 20 cm
2
area
gives the breakdown voltage across a 1 mm gap of 40 kV, whereas
electrodes of the same material of area 1000 cm
2
gives a breakdown
voltage across the same 1 mm gap as 25 kV.
Up to 1 mm gap, the more convex electrodes have higher breakdown
voltage than the more nearly plane electrodes even though at the same
voltage they carried a higher electric field at the surface.
(vi) Temperature
The variation of the breakdown voltage with temperature is very small,
and for nickel and iron electrodes, the strength remains unchanged for
temperatures as high as 500
0
C. Cooling the electrodes to liquid Nitrogen
temperature increases the breakdown voltage.
(vii) Frequency of applied voltage
1 J R Lucas Breakdown of Gaseous Insulation
It is known that a given gap stands a higher impulse voltage than an
alternating voltage, and a higher alternating voltage than a direct voltage.
However, it has been shown that for a small gap (2 mm) there is no
dependence of the breakdown voltage on the frequency in the range 50 Hz
to 50 kHz.
10
-6
10
-5
10
-4
10
-3
Breakdown
Voltage (kV)
1200
1000
800
600
400
200
0
Pressure
(torr)
Figure 1.11 – Breakdown characteristic
2 J R Lucas Breakdown of Gaseous Insulation
(viii) Vacuum Pressure
For small gaps, increasing the degree of vacuum increases the breakdown
voltage until below a certain pressure there is no change.
The vacuum breakdown region is the region in which the breakdown
voltage becomes independent of the nature of the pressure of the gap
between the electrodes.
However, for large gaps (about 200 mm spacing) it is found that below a
certain pressure the breakdown voltage starts decreasing again, so that the
breakdown stress at this stage could in fact be improved by actually
worsening the vacuum.
For example, using a 1.6 mm diameter sphere opposite a plane cathode
and a gap of 200 mm, the characteristics shown in figure 1.11 was
obtained.
2 J R Lucas Breakdown of Gaseous Insulation
The breakdown voltage was constant for pressure less than 5 x 10
- 4
torr,
but rose with increasing pressure to a maximum at 5 x 10
- 4
torr, with
further increase in pressure the voltage fell sharply.
1 J R Lucas Breakdown of Gaseous Insulation
1.2.5 Time lags of Spark breakdown
In high voltage engineering, breakdown under impulse fields is of great
importance.
On the application of a voltage, a certain time elapses before actual
breakdown occurs even though the applied voltage may be much more
than sufficient to cause breakdown.
In considering the time lag observed between the application of a voltage
sufficient to cause breakdown and the actual breakdown the two basic
processes of concern are the appearance of avalanche initiating electrons
and the temporal growth of current after the criterion for static breakdown
is satisfied.
3 J R Lucas Breakdown of Gaseous Insulation

In the case of slowly varying fields, there is usually no difficulty in
finding an initiatory electron from natural sources (ex. cosmic rays,
detachment of gaseous ions etc).
However, for impulses of short duration (around 1 μs), depending on the
gap volume, natural sources may not be sufficient to provide an initiating
→ t ←



Breakdo
wn
Voltage
V

V
s


0
time
(µ s)
Figure 1.12 - Time lag of
impulse breakdown
Impulse
Voltage
Collapse
of
Voltage
2 J R Lucas Breakdown of Gaseous Insulation
electron while the voltage is applied, and in the absence of any other
source, breakdown will not occur.
Figure 1.12 shows the time lag ‘t’ for an impulse voltage waveform.
The time t
s
which elapses between the application of a voltage greater than
or equal to the static breakdown voltage (V
s
) to the spark gap and the
appearance of a suitably placed initiatory electron is called the statistical
time lag of the gap, the appearance being usually statistically distributed.
After such an electron appears, the time t
f
required by the ionisation
processes to generate a current of a magnitude which may be used to
specify breakdown of the gap is known as the formative time lag.
The sum t
f
+ t
s
= t is the total time lag, and is shown in the diagram. The
ratio V/V
s
, which is greater than unity, is called the impulse ratio, and
clearly depends on t
s
+ t
f
and the rate of growth of the applied voltage.
1 J R Lucas Breakdown of Gaseous Insulation
The breakdown can also occur after the peak of the voltage pulse, (i.e. on
the wavetail). Depending on the time lag, the breakdown may occur in
the wavefront or the wavetail of the impulse waveform.
(i) Statistical Time lag t
s
The statistical time lag is the average time required for an electron to
appear in the gap in order that breakdown may be initiated.
If β = rate at which electrons are produced in the gap by external
irradiation
P
1
= probability of an electron appearing in a region of the gap where
it can lead to a spark
P
2
= probability that such an electron appearing in the gap will lead to
a spark
3 J R Lucas Breakdown of Gaseous Insulation
then, the average time lag
P P
1
=
t
2 1
s
β
If the level of irradiation is increased, β increases and therefore t
s
decreases.
Also, with clean cathodes of higher work function β will be smaller for a
given level of illumination producing longer time lags.
The type of irradiation used will be an important factor controlling P
1
, the
probability of an electron appearing in a favourable position to produce
breakdown.
The most favourable position is, of course near the cathode.
If the gap is overvolted, then the probability of producing a current
sufficient to cause breakdown rapidly increases.
3 J R Lucas Breakdown of Gaseous Insulation
P
2
therefore increases with overvoltage and may tend to unity when the
overvoltage is about 10%.
As P
2
→ 1 with increasing overvoltage, t
s
→ 1/βP
1
.
1 J R Lucas Breakdown of Gaseous Insulation
(ii) Formative time lag
After the statistical time lag, it can be assumed that the initiatory electron
is available which will eventually lead to breakdown.
The additional time lag required for the breakdown process to form is the
formative time lag.
An uninterrupted series of avalanches is necessary to produce the requisite
gap current (μA) which leads to breakdown, and the time rate of
development of ionisation will depend on the particular secondary process
operative.
The value of the formative time lag will depend on the various secondary
ionisation processes.
Here again, an increase of the voltage above the static breakdown voltage
will cause a decrease of the formative time lag t
f
.

1 J R Lucas Breakdown of Gaseous Insulation
(iii) Time lag characteristic
Figure 1.13 - Time-lag characteristics
←time lag characteristic
Static
breakdown
3.24 - Single
core
conductor
channel cable
V
2
core
conductor
Breakdown
characteristic
(kV)
3 J R Lucas Breakdown of Gaseous Insulation
The time lag characteristic is the variation of the breakdown voltage with
time of breakdown, and can be defined for a particular waveshape.
The time lag characteristic based on the impulse waveform is shown in
figure 1.13.

A







Breakdo
wn
Voltage

V
A


0
time
(µ s)
Figure 1.14 – Typical Time lag
characteristics breakdown
Rod
gap (ii)

Protected
Transform
er
Rod
gap (i)

Unprotecte
d
1 J R Lucas Breakdown of Gaseous Insulation
The time lag characteristic is important in designing insulation.
If a rod gap is to provide secondary protection to a transformer, then the
breakdown voltage characteristic of rod gap must be less than that of the
transformer at all times (gap i) to protect it from dangerous surge voltages.
This will ensure that the gap will always flashover before the protected
apparatus.
However, with such a rod gap, the gap setting will be low, as the
sharpness of the two characteristics are different.
Thus it is likely that there would be frequent interruptions, even due to the
smallest overvoltages which would in fact cause no harm to the system.
Thus it is usual to have the rod gap characteristic slightly higher (gap ii)
resulting in the intersection of the characteristics as shown.
In such a case, protection will be offered only in the region where the rod
gap characteristic is lower than that of the transformer.
1 J R Lucas Breakdown of Gaseous Insulation
This crossing point is found from experience for a value of voltage which
is highly unlikely to occur.
The other alternative is of course to increase the transformer characteristic
which would increase the cost of the transformer a great deal.
[This decision is something like saying, it is better and cheaper to replace
1 transformer a year due to this decision than have to double the cost of
each of 100 such transformers in the system.]
1.2.6 Corona Discharges
In a uniform electric field, a gradual increase in voltage across a gap
produces a breakdown of the gap in the form of a spark without any
preliminary discharges.
On the other hand, if the field is non–uniform, an increase in voltage will
first cause a localised discharge in the gas to appear at points with the
highest electric field intensity, namely at sharp points or where the
electrodes are curved or on transmission line conductors.
1 J R Lucas Breakdown of Gaseous Insulation
This form of discharge is called a corona discharge and can be observed
as a bluish luminance.
This phenomena is always accompanied by a hissing noise, and the air
surrounding the corona region becomes converted to ozone. Corona is
responsible for considerable power loss in transmission lines and also
gives rise to radio interference.
1.2.6.1 Mechanism of Corona formation on a 2 conductor line
When a gradually increasing voltage is applied across two conductors,
initially nothing will be seen or heard.
As voltage is increased, air surrounding the conductors get ionised, and at
a certain voltage a hissing noise is heard caused by formation of corona.
This voltage is known as the disruptive critical voltage (dcv).
A further increase in the voltage would cause a visible violet glow around
the conductors. This voltage is the visual corona inception voltage.
1 J R Lucas Breakdown of Gaseous Insulation
If the applied voltage is direct, the glow observed will be uniform on the
positive conductor, while the negative conductor will be more patchy and
often accompanied by streamers at rough points.
In the case of alternating voltages, both conductors appear to have a
uniform glow, but when observed stroboscopically the effect is seen to be
similar to the direct voltage case.
If the voltage is further increased, the corona increases and finally spark
over would occur between the two conductors.
If the conductors are placed quite close together, corona formation may
not occur before the spark over occurs.
The condition for this will be considered later.
The formation of corona causes the current waveform in the line, and
hence the voltage drop to be non-sinusoidal.
1 J R Lucas Breakdown of Gaseous Insulation
It also causes a loss of power.
There is always some electrons present in the atmosphere due to cosmic
radiation etc.
When the line voltage is increased, the velocity of the electrons in the
vicinity of the line increases, and the electrons acquire sufficient velocity
to cause ionisation.
To prevent the formation of corona, the working voltage under fair
weather conditions should be kept at least 10% less than the disruptive
critical voltage.
Corona formation may be reduced by increasing the effective radius.
Thus steel cored aluminium has the advantage over hard drawn copper
conductors on account of the larger diameter, other conditions remaining
the same.
1 J R Lucas Breakdown of Gaseous Insulation
The effective conductor diameter can also be increased by the use of
bundle conductors.
Corona acts as a safety valve for lightning surges, by causing a short
circuit.
The advantage of corona in this instance is that it reduces transients by
reducing the effective magnitude of the surge by partially dissipating its
energy due to corona.
The effect of corona on radio reception is a matter of some importance.
The current flowing into a corona discharge contains high-frequency
components.
These cause interference in the immediate vicinity of the line.
As the voltage is gradually increased, the disturbing field makes its
appearance long before corona loss becomes appreciable.
1 J R Lucas Breakdown of Gaseous Insulation
The field has its maximum value under the line and attenuates rapidly
with distance.
The interference fails to about a tenth at 50 m from the axis of the line.
1 J R Lucas Breakdown of Gaseous Insulation
1.2.6.2 Waveform of Corona Current
3 J R Lucas Breakdown of Gaseous Insulation

volta
ge
tDC
V
DC
V
Sine
wave
Applied
voltage
Corona
current
Figure 1.15 Waveform of
corona current
5 J R Lucas Breakdown of Gaseous Insulation
The shunt current in a line is almost purely capacitive under normal
conditions, and leads the applied voltage by 90
0
, and there is no power
loss in the line under no-load conditions.
When the applied voltage is increased and corona is formed, the air is
rendered conducting, and power loss occurs.
The shunt current would no longer be leading the voltage by 90
0
. Thus
the current waveform would consist of two components.
The lossy component would be non-sinusoidal and would occur only
when the disruptive critical voltage is exceeded in either polarity.
This resultant waveform is shown in figure 1.15.
The corona current can be analysed and shown to possess a strong third
harmonic component.
1 J R Lucas Breakdown of Gaseous Insulation
1.2.6.3 Mechanism of corona formation
The stress surrounding the conductor is a maximum at the conductor
surface itself, and decreases rapidly as the distance from the conductor
increases.
Thus when the stress has been raised to critical value immediately
surrounding the conductor, ionisation would commence only in this
region and the air in this region would become conducting.
The effect is to increase the effective conductor diameter while the
voltage remains constant.
This results in two effects.
Firstly, an increase in the effective sharpness of the conductor would
reduce the stress outside this region, and secondly, this would cause a
reduction of the effective spacing between the conductors leading to an
increase in stress.
1 J R Lucas Breakdown of Gaseous Insulation
Depending on which effect is stronger, the stress at increasing distance
can either increase or decrease.
If the stress is made to increase, further ionisation would occur and
flashover is inevitable.
Under ordinary conditions, the breakdown strength of air can be taken as
30 kV/cm.
Corona will of course be affected by the physical state of the atmosphere.
In stormy weather, the number of ions present is generally much more
than normal, and corona will then be formed at a much lower voltage than
in fair weather.
This reduced voltage is generally about 80% of the fair weather voltage.
The condition for stable corona can be analysed as follows.
Figure 1.16 - Electric Stress in two conductor system
d

c
o
r
o
n
a


e
n
v
e
l
o
p
e
xr ξ c
o
n
d
u
c
t
o
r
effective



radius
2 J R Lucas Breakdown of Gaseous Insulation
The electric stress ξ at a distance x from a conductor of radius x, and
separated from the return conductor by a distance d is given by
l x 2
q
.
1
=
0
π
ε
ξ
where q is the charge on each conductor in length l.
Thus the potential V can be determined from V = ∫ ξ dx
dx .
x 2
q
= V
0
r - d
r
ε
π

Since both charges (+q and -q) produce equal potential differences, the
total potential difference between the two conductors is double this value.
Thus the conductor to neutral voltage, which is half the difference would
be equal to this value.


c
o
r
o
n
a


e
n
v
e
l
o
p
e
c
o
n
d
u
c
t
o
r
effective



radius
1 J R Lucas Breakdown of Gaseous Insulation
Thus the conductor to neutral voltage is

,
`

.
|

r
r - d
.
2
q
= V
e
0
log
ε
π
Therefore the electric stress at distance x is given by
r d if
log
log
>>
r
d
x
V
=

r
r - d
x
V
=
e
x
e
x
ξ
ξ

x
and V can both be peak or both rms]


c
o
r
o
n
a


e
n
v
e
l
o
p
e
dddq
b
q
a
q
c
= – q
a

q
b
+
2 J R Lucas Breakdown of Gaseous Insulation
For three phase lines, with equilateral spacing, when V is the voltage to
neutral and d is the equilateral spacing.
Total charge is zero, so that qc = – qa – qb
qa and – qa are a pair similar to the two conductor case.
Similarly is qb and – qb.
These give the solution to the three phase case with equilateral spacing as
the same as the single phase solution.
For air, ξ
max
= 30 kV/cm, so that ξ
rms
= 30/2 = 21.2 kV/cm.
Since there is no electric stress within the conductor, the maximum stress
will occur when x is a minimum, that is at x = r.
Thus if E
0,rms
is the rms value of the disruptive critical voltage to neutral,
r
d
r
E
= 21.2 =
e
rms 0,
rms
log
ξ
1 J R Lucas Breakdown of Gaseous Insulation
When the surface of the conductor is irregular, it is more liable to corona.
Thus an irregularity factor m
0
is introduced to account for this reduction.
Typical values of this factor are
m
0
= 1.0 for smooth polished conductors,
= 0.98 to 0.93 for roughened conductors,
 0.90 for cables of more than 7 strands, and
= 0.87 to 0.83 for 7 strand cables.
Since the corona formation is affected by the mean free path, and hence
by the air density, a correction factor δ is introduced.
This air density correction factor is given by the usual expression, with p
being the pressure expressed in torr and t being the temperature
expressed in
0
C.
1 J R Lucas Breakdown of Gaseous Insulation
t + 273
p 0.386
=
t + 273
20 + 273
.
760
p
= δ
The disruptive critical voltage can then be written as in the following
equation.
E
0,rms
= 21.2 δ m
0
r log
e
(d/r) kV to neutral
1.2.6.4 Visual Corona
Visual corona occurs at a higher voltage than the disruptive critical
voltage.
For the formation of visual corona, a certain amount of ionization, and the
raising of an electron to an excited state are necessary.
1 J R Lucas Breakdown of Gaseous Insulation
The production of light by discharge is not due to ionisation, but due to
excitation, and subsequent giving out of excess energy in the form of light
and other electromagnetic waves.
To obtain the critical voltage for visual corona formation, the disruptive
critical voltage has to be multiplied by a factor dependant on the air
density and the conductor radius.
Further the value of the irregularity factor is found to be different.
The empirical formula for the formation of visual corona is
r
d
.
r
0.3
+ 1 r
m
21.2 =
E
e
v rms v,
log
]
]
]

δ
δ
The values of the irregularity factor m
v
for visual corona is given by
m
v
= 1.0 for smooth conductors,
3 J R Lucas Breakdown of Gaseous Insulation
= 0.72 for local corona on stranded wires (patches)
= 0.82 for decided corona on stranded wires (all over the wire)
1 J R Lucas Breakdown of Gaseous Insulation
1.2.6.5 Stable Corona formation
Consider two conductors, just on the limit of corona formation.
Assume that there is a thin layer Δr of ionised air around each conductor,
so that the effective radius becomes (r + Δr).
The change in electric stress Δξ due to this layer can be determined using
differentiation. Thus
)
r
d
r (
r . )
r
d
- 1 ( E
= r .
r
d -
.
d
r
. r +
r
d
.
)
r
d
r (
E -
=

r
d
r
E
=
2
e
e
2
e
2
e
e
log
log
log
log
log


]
]
]
]
]
]

,
`

.
|

,
`

.
|
∆ ∆
ξ
ξ
3 J R Lucas Breakdown of Gaseous Insulation
When log
e
> 1, the above expression is negative. i.e. d/r > e (=2.718)
Under this condition, the effective increase in diameter lowers the electric
stress and no further stress increase is formed, and corona is stable.
If on the other hand, d/r < e, then the effective increase in the diameter
raises the electric stress, and this causes a further ionisation and a further
increase in radius, and finally leads to flash-over.
In practice, the effective limiting value of d/r is about 15 and not e
(=2.718). For normal transmission lines, the ratio d/r is very much greater
than 15 and hence stable corona always occurs before flashover.
For example, in a 132 kV line with a spacing of 4 m and a radius of
conductor of 16 mm, the ratio has a value d/r = 4/16x10
-3
= 250 >> 15.
1 J R Lucas Breakdown of Gaseous Insulation
1.2.6.6 Power Loss due to Corona
The formation of corona is associated with a loss of power.
This loss will have a small effect on the efficiency of the line, but will not
be of sufficient importance to have any appreciable effect on the voltage
regulation.
The more important effect is the radio interference.
The power loss due to corona is proportional to the square of the
difference between the line-to-neutral voltage of the line and the
disruptive critical voltage of the line.
It is given by the empirical formula
e kW/km/phas
10
. )
E
- E ( .
d
r
. ) 25 + f ( .
243
=
P
5 -
2
rms 0, c
δ
where E
0,rms
= disruptive critical voltage (kV)
3 J R Lucas Breakdown of Gaseous Insulation
f = frequency of supply (Hz)
E = Phase Voltage (line to neutral) (kV)
For storm weather conditions, the disruptive critical voltage is to be taken
as 80% of disruptive critical voltage under fair weather conditions.
Example
Determine the Disruptive Critical Voltage, the Visual Corona inception
voltage, and the power loss in the line due to corona, both under fair
weather conditions as well as stormy weather conditions for a 100 km
long 3 phase, 132 kV line consisting of conductors of diameter 1.54 cm,
arranged in an equilateral triangle configuration with 3 m spacing.
The temperature of the surroundings is 40
0
C and the pressure is 750 torr.
The operating frequency is 50 Hz. [The irregularity factors may be taken
as m
o
= 0.83, m
v
= 0.72]
The Air density correction factor δ is given by
3 J R Lucas Breakdown of Gaseous Insulation

5 0. = .
760
750
=
t + 273
p 0.386
=
t + 273
20 + 273
.
760
p
=
92
313
293
δ
δ

Disruptive critical voltage = 21.2 δ m
0
r log
e
(d/r) kV to neutral
= 21.2 x 0.925 x 0.83 x 0.77 x log
e
(3/0.0077)
= 74.76 kV
132/3 = 76.21 kV. Thus corona inception will normally occur.
Similarly, Visual corona voltage
= 21.2 x 0.925 x 0.72 x 0.77 x log
e
(3/0.0077) x [ 1 + 0.3/(0.895 x 0.77)
½
]
= 88.3 kV. Thus visual corona will normally not occur.
1 J R Lucas Breakdown of Gaseous Insulation
Power loss under fair weather condition is given by
kW/phase 1 . 2 = kW/phase 76 . 74
77
925 .
100 .
10
.
3
132
.
3
0.00
. 25) + (50 .
0
243
=
P
5 -
2
c

,
`

.
|

Similarly, power loss under stormy weather condition is given by
kW/phase 268 =
kW/phase 8 . 0 * 76 . 74
77
925 .
100 .
10
.
3
132
.
3
0.00
. 25) + (50 .
0
243
=
P
5 -
2
c

,
`

.
|

Also corona inception voltage = 74.76*0.8 = 59.8 kV, and
Visual corona inception voltage = 88.3*0.8 = 70.6 kV
It is seen that although visual corona does not occur under normal
conditions, it would occur under stormy conditions.
2
Breakdown of
Liquid and Solid Insulation
Breakdown of Liquid and Solid Insulation J R Lucas
2.1Breakdown in Liquids
In highly purified liquid dielectrics, breakdown is controlled by phenomena
similar to those for gasses and electric strength is high (of order of 1 MV/cm).
Unfortunately, liquids are easily contaminated, and may contain solids, other
liquids in suspension and dissolved gasses.
Effect of these impurities is relatively small for short duration pulses (10 μs).
However, if the voltage is applied continuously, the solid impurities line up at
right angles to equipotentials, and distort the field so that breakdown occurs at
relatively low voltage.
The line up of particles is a fairly slow process, and is unlikely to affect the
strength on voltages lasting for less than 1 ms.
Under the action of the electric field, dissolved gasses may come out of solution,
forming a bubble.
3
Breakdown of Liquid and Solid Insulation J R Lucas
The gas in the bubble has a lower strength than the liquid, so that more gas is
produced and the bubble grows, ultimately causing breakdown.
Because, of the tendency to become contaminated, liquids are not usually used
alone above 100 kV/cm in continuously energised equipment.
However, they are used at much higher tresses (up to 1 MV/cm) in conjunction
with solids, which can be made to act as barriers, preventing the line-up of solid
impurities and localising of any bubbles which may form.
The main function of the liquid in such arrangements is to fill up the voids.
4
Breakdown of Liquid and Solid Insulation J R Lucas
2.1.1 Breakdown of Commercial liquids
When a difference of potential is applied to a pair of electrodes immersed in an
insulating liquid, a small conduction current is first observed.
If the voltage is raised continuously, at a critical voltage a spark passes between
the electrodes.
The passage of a spark through a liquid involves the following.
(a) flow of a relatively large quantity of electricity, determined by the
characteristics of the circuit,
(b) a bright luminous path from electrode to electrode,
(c) the evolution of bubbles of gas and the formation of solid products of
decomposition (if the liquid is of requisite chemical nature)
(d) formation of small pits on the electrodes,
(e) an impulsive pressure through the liquid with an accompanying explosive
sound.
5
Breakdown of Liquid and Solid Insulation J R Lucas
Tests on highly purified transformer oil show that
(a) breakdown strength has a small but definite dependence on electrode
material,
(b) breakdown strength decreases with increase in electrode spacing,
(c) breakdown strength is independent of hydrostatic pressure for degassed oil,
but increases with pressure if oil contains gases like nitrogen or oxygen in
solution.
In the case of commercial insulating liquid, which may not be subjected to very
elaborate purifying treatment, the breakdown strength will depend more upon
the nature of impurities it contains than upon the nature of the liquid itself.
These impurities which lead to the breakdown of commercial liquids below their
intrinsic strength, can be divided into the following 3 categories.
6
Breakdown of Liquid and Solid Insulation J R Lucas
(a) Impurities which have a breakdown strength lower than that of the liquid
itself (ex: bubbles of gas).
Breakdown of the impurities may trigger off the total breakdown of the
liquid.
(b) Impurities which are unstable in the electric field (ex: globules of water).
Instability of the impurity can result in a low resistance bridge across the
electrodes and in total breakdown.
(c) Impurities which result in local enhancement of electric field in a liquid (ex:
conducting particles).
The enhanced field may cause local breakdown and therefore initiate
complete breakdown.
These will be considered in turn in the following sections.
7
Breakdown of Liquid and Solid Insulation J R Lucas
2.1.2 Breakdown due to gaseous inclusions
Gas or vapour bubbles may exist in impure liquid dielectrics, either formed from
dissolved gasses, temperature and pressure variations, or other causes.
The electric field E
b
in a gas bubble which is immersed in a liquid of
permittivity ε
1
is given by
E
1 + 2
3
=
E 0
1
1
b
ε
ε
where E
0
is the field in the liquid in the absence of the bubble.
The electrostatic forces on the bubble cause it to get elongated in the direction of
the electric field.
The elongation continues, when sufficient electric field is applied, and at a
critical length the gas inside the bubble (which has a lower breakdown strength)
8
Breakdown of Liquid and Solid Insulation J R Lucas
breaks down. This discharge causes decomposition of the liquid molecules and
leads to total breakdown.
2.1.3 Breakdown due to liquid globules
If an insulating liquid contains in suspension a globule of another liquid, then
breakdown can result from instability of the globule in the electric field.
Consider a spherical globule of liquid of permittivity ε
2
immersed in a liquid
dielectric of permittivity ε
1
.
When it is subjected to an electric field between parallel electrodes, the field
inside the globule would be given by
E
+ 2
3
= E
0
2 1
1
ε ε
ε
where E
0
is the field in the liquid in the absence of the globule.
The electrostatic forces cause the globule to elongate and take the shape of a
prolate spheroid (i.e. an elongated spheroid).
9
Breakdown of Liquid and Solid Insulation J R Lucas
As the field is increased, the globule elongates so that the ratio γ of the longer to
the shorter diameter of the spheroid increases.
For the same field E, the ratio γ is a function of ε
2

1
.
10
Figure 2.1 - Variation of ratio of
diameters of spheroid
ε
2
/
ε
1
=

ε
2

1
> 20
ε
2

1
< 20
ε
2

1
= 0.5
Electric
stress E
E
c
rit
ratio γ
of
diamete
rs of
spheroi
d
Breakdown of Liquid and Solid Insulation J R Lucas
When ε
2
>> ε
1
(generally when ε
2

1
> 20), and the field exceeds a critical value,
no stable shape exists, and the globule keeps on elongating eventually causing
bridging of the electrodes, and breakdown of the gap.
When ε
2

1
>> 20, the critical field at which the globule becomes unstable no
longer depends on the ratio, and is given by E
crit.
kV/cm
R
1.542 =
E
1
crit

,
`

.
|
ε
σ
2
1
where σ = surface tension of the globule (N/m)
ε
1
= relative permittivity of the insulating liquid
R = initial radius of globule (m).
Example 2.1
11
Breakdown of Liquid and Solid Insulation J R Lucas
For a droplet of water (R = 1 μm , ε
2
= 90) in an insulating oil (ε
1
= 2); ε
2
>> ε
1
.
Also σ = 0.043 N/m.
Thus E
crit
= 1.542 (0.043/10
–6
x 2)
½
kV/cm = 226 kV/cm.
= 0.226 MV/cm
Thus we see that a droplet of water even as small as 1 μm in radius (quite
unobservable) can greatly reduce the breakdown strength of the liquid dielectric.
In fact, a globule of water of radius of only 0.05 μm, which is quite
unobservable, will be disrupted at a value of about 1 MV/cm which is the
breakdown strength of the pure liquid.
Thus even submicroscopic sources of water, such as condensed breakdown
products, or hygroscopic solid impurities, may greatly influence breakdown
conditions.
A globule which is unstable at an applied value of field elongates rapidly, and
then electrode gap breakdown channels develop at the end of the globule.
12
Breakdown of Liquid and Solid Insulation J R Lucas
Propagation of the channels result in total breakdown.
13
Breakdown of Liquid and Solid Insulation J R Lucas
2.1.4 Breakdown due to solid particles
In commercial liquids, solid impurities cannot be avoided and will be present as
fibres or as dispersed solid particles.
If the impurity is considered to be a spherical particle of permittivity ε
2
and is
present in a liquid dielectric of permittivity ε
1
, it will experience a force
E
2 +
) - (

r
2
1
= F
2
1 2
1 2
0
3

ε ε
ε ε
ε
where E = applied field, r = radius of particle.
Generally ε
2
> ε
1
, so that the force would move the particle towards the regions
of stronger field.
Particles will continue to move in this way and will line up in the direction of
the field.
14
Breakdown of Liquid and Solid Insulation J R Lucas
A stable chain of particles would be produced, which at a critical length may
cause breakdown.
Because of the tendency to become contaminated, liquids are seldom used alone
above 100 kV/cm in continuously energised equipment.
However they may be used up to 1 MV/cm in conjunction with solids which can
be made to act as barriers, preventing the line–up of solid impurities and
localising bubbles which may form.
15
Breakdown of Liquid and Solid Insulation J R Lucas
2.1.5 Purification of a liquid for testing
(a) Removal of dust
Small dust particles can become charged and cause local stresses which can
initiate breakdown.
They can also coalesce to form conducting bridges between electrodes.
Careful filtration can remove dust particles greater in size than 1 μm.
The strength of the liquid then increases and greater stability is achieved.
(b) Removal of dissolved gasses
Liquid insulation will normally contain dissolved gas in small but significant
amounts.
Some gases such as Nitrogen and Hydrogen do not appear to upset the
electrical properties to a great extent, but oxygen and carbon dioxide can
cause the strength to change significantly.
Thus it necessary to control the amount of gases present. This is done by
distillation and degassing.
16
Breakdown of Liquid and Solid Insulation J R Lucas
(c) Removal of ionic impurities
Ionic impurities in the liquid (particularly residual water which easily
dissociates) leads to abnormal conductivity and heating of the liquid.
Water can be removed by drying agents, vacuum drying, and by freezing out
in low temperature distillation.
For measurements on liquid dielectrics, where test cells are small, electrode
preparation is much more critical than it is for measurements on gases or solids.
Not only is the surface smoothness important, but surface films, particularly
oxides can have a marked influence on the strength.
17
Breakdown of Liquid and Solid Insulation J R Lucas
2.2Breakdown of Solid Insulating Materials
In solid dielectrics, highly purified and free of imperfections, the breakdown
strength is high, of the order of 10 MV/cm.
The highest breakdown strength obtained under carefully controlled conditions
is known as the "intrinsic strength" of the dielectric.
Dielectrics usually fail at stresses well below the intrinsic strength due usually to
one of the following causes.
(a) electro–mechanical breakdown
(b) breakdown due to internal discharges
(c) surface breakdown (tracking and erosion)
(d) thermal breakdown
(e) electro–chemical breakdown
(f) chemical deterioration
These will now be considered in the following sections.
18
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.1 Electro–mechanical breakdown–
When an electric field is applied to a dielectric between two electrodes, a
mechanical force will be exerted on the dielectric due to the force of attraction
between the surface charges.
This compression decreases the dielectric thickness thus increasing the effective
stress. This is shown in figure 2.2.
19
Figure 2.2 - Process
of breakdown
absence of field
with applied field
d
o
d
Breakdown of Liquid and Solid Insulation J R Lucas
Compressive force P
c
= ½ D E = ½ ε
o
ε
r
V
2
/d
2
,
and From Hooke's Law for large strains, P
c
= Y ln (d
o
/d)
At equilibrium, equating forces gives the equation,
d
d
d

Y 2
=
V
o 2
r o
2
ln
ε ε
By differentiating with respect to d, it is seen that the system becomes
unstable when ln (d
o
/d) > ½ or d < 0.6 d
o
.
Thus when the field is increased, the thickness of the material decreases.
At the field when d < 0.6 d
o
, any further increase in the field would cause the
mechanical collapse of the dielectric.
The apparent stress (V/d
o
) at which this collapse occurs is thus given by the
equation
20
Breakdown of Liquid and Solid Insulation J R Lucas
]
]
]

Y
0.6 =
E
r o
2
1
a
ε ε
21

void V
v
C
v
C
s
C
p
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.2 Breakdown due to internal discharges
Solid insulating materials sometimes contain voids or cavities in the medium or
boundaries between the dielectric and the electrodes.
These voids have a dielectric constant of unity and a lower dielectric strength.
Hence the electric field strength in the voids is higher than that across the
dielectric.
Thus even under normal working voltages, the field in the voids may exceed
their breakdown value and breakdown may occur.
The mechanism can be explained by considering the following equivalent circuit
of the dielectric with the void, shown in figure 2.3.
When the voltage V
v
across the void exceeds the critical voltage V
c
, a discharge
is initiated and the voltage collapses.
The discharge extinguishes very rapidly (say 0.1 μs). The voltage across the
void again builds up and the discharges recur.
22
F
i
g
u
r
e
2
.
3
-
E
q
u
i
v
a
l
e
n
t
c
ir
c
u
it
o
f
d
i
e
l
e
c
tr
i
c
w
it
h
v
o
i
d
Breakdown of Liquid and Solid Insulation J R Lucas
The number and frequency of the discharges will depend on the applied voltage.
The voltage and current waveforms (exaggerated for clarity) are shown in figure
2.4.
In each of the discharges, there will be heat dissipated in the voids which will
cause carbonization of the surface of the voids and erosion of the material.
The gradual erosion of the material and consequent reduction in the thickness of
the insulating material eventually leads to breakdown.
23
F
i
g
u
r
e
2
.
3
-
E
q
u
i
v
a
l
e
n
t
c
ir
c
u
it
o
f
d
i
e
l
e
c
tr
i
c
w
it
h
v
o
i
d
Breakdown of Liquid and Solid Insulation J R Lucas
Breakdown by this process is slow and may occur in a few days or may take a
few years.
Deterioration due to internal discharges
24
F
i
g
u
r
e
2
.
3
-
E
q
u
i
v
a
l
e
n
t
c
ir
c
u
it
o
f
d
i
e
l
e
c
tr
i
c
w
it
h
v
o
i
d
voltage across void if no
discharges occurred
applied
voltage
vV
C

V
C
tit
Figure 2.4 -
Internal discharges
Breakdown of Liquid and Solid Insulation J R Lucas
In organic liquid-solid dielectrics, internal discharges produce gradual
deterioration because of
(a) disintegration of the solid dielectric under the bombardment of electrons set
free by the discharges
(b) chemical action on the dielectric of the products of ionization of the gas
(c) high temperatures in the region of the discharges.
All voids in the dielectric can be removed by careful impregnation and this
results in an increase in the discharge inception stress E
i
. The final value E
i
then
depends on electrical processes which lead to gas formation.
In oil impregnated paper these are
(a) decomposition of moisture in paper
(b) local electrical breakdown of the oil.
25
Breakdown of Liquid and Solid Insulation J R Lucas
Stress at which gas is evolved from paper containing appreciable quantities of
moisture can be less than 10 V/μm, but increases continuously with increasing
dryness and can be higher than 100 V/μm when the paper is thoroughly dry.
Except in very dry conditions, the gas first formed arises from electrochemical
decomposition of water held in the paper.
When a gas bubble is formed in an oil-paper dielectric at the discharge inception
stress E
i
, discharges in the bubble decompose the molecules of the oil, resulting
in further gas formation and a rapid growth of the bubble.
As long as the bubble remains in the dielectric, the inception stress E
i
is low,
often lower than the rated stress, but resting the dielectric long enough for the
gas to dissolve in the oil restores the initial high discharge inception stress.
Although on resting E
i
improves, permanent damage has been caused by the
discharges and this manifests itself in an increase of loss angle and is due to the
formation of ions by the discharges.
Also, due to the discharges, widespread carbonization occur.
26
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.3 Surface Breakdown
Surface flashover
Surface flashover is a breakdown of the medium in which the solid is immersed.
The role of the solid dielectric is only to distort the field so that the electric
strength of the gas is exceeded.
If a piece of solid insulation is inserted in a gas so that the solid surface is
perpendicular to the equipotentials at all points, then the voltage gradient is not
affected by the solid insulation.
An example of this is a cylindrical insulator placed in the direction of a uniform
field.
Field intensification results if solid insulation departs even in detail from the
cylindrical shape.
In particular if the edges are chipped, or if the ends of the cylinder are not quite
perpendicular to the axis, then an air gap exists next to the electrode, and the
stress can reach up to ε
r
times the mean stress in the gap.
27
Breakdown of Liquid and Solid Insulation J R Lucas

r
is the dielectric constant of the cylinder].
Discharge may therefore occur at a voltage approaching 1/ε
r
times the
breakdown voltage in the absence of the cylinder, and these discharges can
precipitate a breakdown.
The three essential components of the surface flashover phenomena are
(a) the presence of a conducting film across the surface of the insulation
(b) a mechanism whereby the leakage current through the conducting film is
interrupted with the production of sparks,
(c) degradation of the insulation must be caused by the sparks.
The conducting film is usually moisture from the atmosphere absorbed by some
form of contamination.
Moisture is not essential as a conducting path can also arise from metal dust due
to wear and tear of moving parts.
28
Breakdown of Liquid and Solid Insulation J R Lucas
Sparks are drawn between moisture films, separated by drying of the surface due
to heating effect of leakage current, which act as extensions to the electrodes.
For a discharge to occur, there must be a voltage at least equal to the Paschen
minimum for the particular state of the gas.
For example, Paschen minimum in air at N.T.P it is 380 V, whereas tracking can
occur at well below 100 V.
It does not depend on gaseous breakdown.] Degradation of the insulation is
almost exclusively the result of heat from the sparks, and this heat either
carbonises if tracking is to occur, or volatilises if erosion is to occur.
Carbonization results in a permanent extension of the electrodes and usually
takes the form of a dendritic growth.
Increase of creepage path during design will prevent tracking, but in most
practical cases, moisture films can eliminate the designed creepage path.
29
Breakdown of Liquid and Solid Insulation J R Lucas
Tracking
Tracking is the formation of a permanent conducting path across a surface of the
insulation, and in most cases the conduction (carbon path) results from
degradation of the insulation itself leading to a bridge between the electrodes.
The insulating material must be organic in nature for tracking to occur.
Erosion
In a surface discharge, if the products of decomposition are volatile and there is
no residual conducting carbon on the surface, the process is simply one of
pitting.
This is erosion, which again occurs in organic materials.
If surface discharges are likely to occur, it is preferable to use materials with
erosion properties rather than tracking properties, as tracking makes insulation
immediately completely ineffective, whereas erosion only weakens the material
but allows operation until replacement can be made later.
30
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.4 Thermal Breakdown
Heat is generated continuously in electrically stressed insulation by dielectric
losses, which is transferred to the surrounding medium by conduction through
the solid dielectric and by radiation from its outer surfaces.
If the heat generated exceeds the heat lost to the surroundings, the temperature
of the insulation increases.
The power dissipated in the dielectric can be calculated as follows.
Uniform direct stress
Power dissipated/volume = ξ
2
/ρ W/m
3

where ξ = uniform direct stress V/m
ρ = resistivity of insulation Ω m
31
Breakdown of Liquid and Solid Insulation J R Lucas
Uniform alternating stress
Power dissipated P = V . I cos φ = V . VCω tan δ
where V = applied voltage V
ω = supply frequency Hz
C = dielectric capacitance F = A ε
r
ε
0
/ d
δ = loss angle rad
ε = dielectric constant
Therefore P = V
2
(A ε
r
ε
0
/ d) ω tan δ
ξ = alternating stress V/m
Re-arranging terms gives the result
P = (V/d)
2
. ε
r
ε
0
. 2 π f . tan δ . A . d
32
Breakdown of Liquid and Solid Insulation J R Lucas
Since A.d is the volume of the dielectric, and V/d is the uniform applied stress,
Power dissipated/volume= ξ
2
ε
r
ε
0
2 π f tan δ W/m
3

= 2 π x 8.854 x 10
-12
ξ
2
ε
r
f tan δ W/m
3
= 5.563 x 10
-11
ξ
2
ε
r
f tan δ x 10
10
W/m
3
if ξ is in kV/cm
Power dissipated/volume= 0.556 ξ
2
f ε
r
tan δ W/m
3
with ξ in kV/cm
The simplest case is where the loss of heat by cooling is linearly related to the
temperature rise above surroundings, and the heat generated is independent of
temperature. (i.e. the resistivity and the loss angle do not vary with
temperature).
Heat lost = k (θ - θ
0
), where θ = ambient temperature
33
Breakdown of Liquid and Solid Insulation J R Lucas
Equilibrium will be reached at a temperature θ
1
where the heat generated is
equal to the heat lost to the surroundings, as shown in figure 2.5.
34
Figure 2.5
Thermal
breakdown
θ
0
θ
1
tempe
rature
h
ea
t
heat
generat
ed
heat
lost
Breakdown of Liquid and Solid Insulation J R Lucas
In practice, although the heat lost may be considered somewhat linear, the heat
generated increases rapidly with temperature, and at certain values of electric
field no stable state exists where the heat lost is equal to the heat generated so
that the material breaks down thermally.
Figure 2.6 shows the variation of heat generated by a device for 2 different
applied fields and the heat lost from the device with temperature.
35
Figure 2.6 Thermal breakdown
with non-linear heat generation
θ
0
θ
A
tempe
rature
h
ea
t
heat
generat
ed
heat
lost
θ
B
E
1
E
2
Breakdown of Liquid and Solid Insulation J R Lucas
The rapid increase is due to the fact that with rise in temperature, the loss angle
of the dielectric increases in accordance with an exponential law (loss ∝ e
-A/T
,
where T is the absolute temperature).
For the field E
2
, a stable temperature θ
A
exists (provided the temperature is not
allowed to reach θ
B
).
For the field E
1
, the heat generated is always greater than the heat lost so that the
temperature would keep increasing until breakdown occurs.
The maximum voltage a given insulating material can withstand cannot be
increased indefinitely simply by increasing its thickness.
Owing to thermal effects, there is an upper limit of voltage V
θ
, beyond which it
is not possible to go without thermal instability.
This is because with thick insulation, the internal temperature is little affected by
the surface conditions.
Usually, in the practical use of insulating materials, V
θ
is a limiting factor only
for high-temperature operation, or at high frequency failures.
36
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.5 Electro–chemical Breakdown–
Since no insulant is completely free of ions, a leakage current will flow when an
electric field is applied.
The ions may arise from dissociation of impurities or from slight ionisations of
the insulating material itself.
When these ions reach the electrodes, reactions occur in accordance with
Faraday's law of electrolysis, but on a much smaller scale.
The insulation and the electrode metal may be attacked, gas may be evolved or
substance may be deposited on the electrodes.
The products of the electrode reaction may be chemically or electrically harmful
and in some cases can lead to rapid failure of the insulation.
The reactions are much slower than in normal electrolytic processes due to the
much smaller currents.
37
Breakdown of Liquid and Solid Insulation J R Lucas
The products of the reactions may be electrically and chemically harmful
because the insulation and electrodes may be attacked, and because harmful
gases may be evolved.
Typically a 1 μF paper capacitor operating at 1 kV at room temperature would
require 2 to 3 years to generate 1 cm
3
hydrogen.
At elevated temperatures, the products of electrolysis would be formed much
more rapidly.
Also since impurities give rise to an increase in the ion concentration, care must
be taken to prevent contamination during manufacture.
The rate of electrolysis is much greater with direct stress than with alternating
stress.
This is due to the fact that the reactions may be wholly or partially reversed
when the polarity changes and the extent of reaction depends on the reaction rate
and the time for diffusion of the reaction products away from the electrodes as
well as on the nature of the reaction products.
38
Breakdown of Liquid and Solid Insulation J R Lucas
However at power frequency, electrochemical effects can be serious and are
often responsible for long-term failure of insulation.
The most frequent source of ions is ionizable impurities in the insulation.
Thus contamination of insulation during manufacture and during assembly into
equipment must be avoided with great care.
Also, contamination in polar insulating materials should be avoided with still
greater care because of the greater degree of dissociation of ionic substance in
solution.
The long term lives of capacitors containing chlorinated impregnants under
direct stress may be greatly extended by adding small quantities of certain
stabilizers, which are hydrogen acceptors and act as depolarizers at the cathode.
Hydrogen ions discharged at the cathode readily react with the stabilizer rather
than with the impregnant, a more difficult chemical process.
39
Breakdown of Liquid and Solid Insulation J R Lucas
In the absence of the stabilizer, the hydrogen reacts with the chlorine of the
impregnant to produce hydrochloric acid, and rapid deterioration occurs due to
attack of the acid on the electrodes and cellulose.
The extension of the life caused by the stabilizers is proportional to the amount
of stabilizer added. For example, with 2% of the stabilizer Azobenzene, mean
life may be extended 50 times.
40
Breakdown of Liquid and Solid Insulation J R Lucas
2.2.6 Chemical Deterioration
Progressive chemical degradation of insulating materials can occur in the
absence of electric stress from a number of causes.
Chemical Instability
Many insulating materials, especially organic materials, show chemical
instability.
Such chemical changes may result from spontaneous breakdown of the structure
of the material.
Under normal operating conditions, this process is very slow, but the process is
strongly temperature dependant.
The logarithm of the life t of paper insulation can be expressed as an inverse
function of the absolute temperature θ.
41
Breakdown of Liquid and Solid Insulation J R Lucas
log
10
t = A/θ + B where A & B are constants
If t is expressed in weeks, for vacuum dried paper immersed in oil in
contact with Nitrogen, the constants have values A = 7000 and
B = - 16.0.
42
Figure 2.7 - Dependence of
life of paper on temperature
(
o
C
)
Life of
Paper
(weeks
)
1
0
0
×
0.1%
water
1
0
10.
1
0.
01
0.0
01
0.0
02
0.0
022
0.0
024
0.0
026
0.0
028
2
0
0
o
1
5
0
o
1
0
0
o
8
0
o
(K
−1
)
Breakdown of Liquid and Solid Insulation J R Lucas
In the presence of oxygen or moisture, the life of the insulation decreases much
more rapidly.
With increase in amount of moisture present, B decreases so that the life of the
paper also decreases.
With about 0.1% moisture present, B decreases by as much as 0.8, so that t
decreases by a factor of about 6.
This means that presence of about 0.1% moisture reduces the life of the
insulation by as much as 6 times. Figure 2.7 shows the variation.
Hydrolysis
When moisture or water vapour is present on the surface of a solid dielectric,
hydrolysis occurs and materials lose their electrical and mechanical properties.
Electrical properties of materials such as paper, cotton tape, and other cellulose
materials deteriorate very rapidly due to hydrolysis.
43
Breakdown of Liquid and Solid Insulation J R Lucas
Polyethylene film may lose its mechanical strength in a few days if kept at 100
% relative humidity.
Oxidation
In the presence of air or oxygen, especially ozone, materials such as rubber and
polyethylene undergo oxidation giving rise to surface cracks, particularly if
stretched and exposed to light.
Polythene also oxidises in strong day light unless protected by an opaque filler.
Other processes
Progressive chemical degradation of insulating materials can also occur due to a
variety of processes such as,
incompatibility of materials (ex: rubber ages more rapidly at elevated
temperatures in the presence of copper, and cellulose degrades much more
rapidly in the presence of traces of acidic substances), and
44
Breakdown of Liquid and Solid Insulation J R Lucas
leaching (washing out of a soluble constituent) of chemically active substances
(ex: glass fabrics made from glasses of high sodium content lose their strength
rapidly due to leaching of sodium to the surface of the fibres and the subsequent
chemical attack of the strong alkali on the glass surface).
2.3Breakdown of Composite Insulation
Almost no complete electrical insulation consists of one insulating phase.
Usually more than one insulating material will be involved, either in series,
parallel or both.
The simplest form of composite insulation system consists of 2 layers of the
same material.
In this case advantage is taken of the fact that two thin sheets have a higher
electric strength than a single sheet of the same total thickness.
In other cases, composite dielectrics occur either due to design considerations
(ex: paper with an impregnating liquid) or due to practical difficulties of
fabrication (ex: air in parallel with solid insulation).
45
Breakdown of Liquid and Solid Insulation J R Lucas
In certain cases, the behaviour of the composite insulation could be predicted
from the behaviour of the components.
But in most cases, the system as whole has to be considered.
The following considerations determine the performance of the system as a
whole.
(i) The stress distribution at different parts of the insulation system is
distorted due to the component dielectric constants and conductivities,
(ii) the breakdown characteristics at the surface are affected by the insulation
boundaries of various components,
(iii) the internal or partial discharge products of one component invariably
affect the other components in the system, and
(iv) the chemical ageing products of one component also affect the
performance of other components in the system.
46
Breakdown of Liquid and Solid Insulation J R Lucas
47
Breakdown of Liquid and Solid Insulation J R Lucas
2.3.1 Matching dielectric constants
When composite insulation has components with different dielectric constants,
utilisation of the materials may be impaired.
This is especially true in the oil/transformerboard dielectric.
This is because the oil has a lower dielectric constant and lower dielectric
strength compared to that of transformerboard.
Since the dielectrics are in series,
V .
d
+
d
=
d
V
=
1 2 2 1
2
1
1
1
ε ε
ε
ξ
48
Breakdown of Liquid and Solid Insulation J R Lucas
V .
d
+
d
=
d
V
=
1 2 2 1
1
2
2
2
ε ε
ε
ξ
49
Figure 2.8 -
Composite
Dielectric
V
1
/
2
V
1
/
2
d
1
/
2
d
1
/
2
d
2
5
0
0
Breakdown of Liquid and Solid Insulation J R Lucas

d
+
d
d
=
V
+
V
V
=
V
V
V
+
V
= V

d
d
=
A
d
.
d
A
=
C
C
=
V
V
1 2 2 1
1 2
2 1
1 1
2 1
2 1
1 2
1
1
2
2
1
2
2
1
ε ε
ε
ε
ε
ε
ε
Example
A transformer oil having a dielectric constant of 2.2 and a dielectric strength of
25 kV/mm, is used as an insulation in a of spacing 8 mm.
Determine the maximum applicable voltage. A barrier of thickness 3 mm of
transfomerboard with a higher dielectric strength of 50 kV/mm (dielectric
constant 4.4) is used in this space to increase the strength.
Does the transformerboard serve this purpose in this case ?
With only transformer oil, the maximum applicable voltage is given by
50
Breakdown of Liquid and Solid Insulation J R Lucas
V = 25× 8 = 800 kV
If a barrier of thickness 3 mm is placed in the space with the oil, the maximum
applicable voltage is given by
kV 350 =
2
14 50
= V
V
6 4.4 + 2 2.2
4.4
= 50
×
×
× ×
It can be seen that the maximum applicable voltage in fact reduces below that of
only oil.
It is thus important, when barriers have to be used, to match the permittivities of
the component insulations.
Thus great gains could be achieved if a transformerboard with a dielectric
constant of 2.6 could be used instead of one with 4.4.
51
Breakdown of Liquid and Solid Insulation J R Lucas
52
3
High Voltage Cables
54 J.R.Lucas
High Voltage Cables
3.0High Voltage Cables
High Voltage Cables are used when underground transmission is required.
These cables are laid in ducts or may be buried in the ground.
Unlike in overhead lines, air does not form part of the insulation, and the
conductor must be completely insulated.
Thus cables are much more costly than overhead lines.
Also, unlike for overhead lines where tappings can easily given, cables
must be connected through cable boxes which provide the necessary
insulation for the joint.
Cables have a much lower inductance than overhead lines due to the lower
spacing between conductor and earth, but have a correspondingly higher
capacitance, and hence a much higher charging current.
High voltage cables are generally single cored, and hence have their separate
insulation and mechanical protection by sheaths.
In the older paper insulated cables, the sheath was of extruded lead.
55 J.R.Lucas
High Voltage Cables
Figure 3.1 shows three such cables, as usually laid out.
The presence of the sheath introduces certain difficulties as currents are
induced in the sheath as well.
This is due to fact that the sheaths of the conductors cross the magnetic
fields set up by the conductor currents.
At all points along the cable, the magnetic field is not the same.
Hence different voltages are induced at different points on the sheath.
This causes eddy currents to flow in the sheaths.
These eddy currents depend mainly on
(a) the frequency of operation,
(b) the distance between cables,
(c) the mean radius of the sheath, and
Figure 3.1 - Layout of three,
single-core cables
56 J.R.Lucas
conducto
r loss
dielectri
c loss
sheath
loss &
interheat
h loss
Heat
generate
d
High Voltage Cables
(d) the resistivity of the sheath material.
3.1Power loss in the Cable

Power loss in the cable can occur due to a variety of reasons (Figure 3.2).
They may be caused by the
conductor loss (also sometimes called the copper loss on account of the
fact that conductors were mainly made out of copper) due to the
conductor current passing through the resistance of the conductor,
dielectric losses caused by the voltage across the insulation,
Figure 3.2 - Heat Transfer
in a cable due to losses
57 J.R.Lucas
High Voltage Cables
sheath losses caused by the induced currents in the sheath, and
intersheath losses caused by circulating currents in loops formed between
sheaths of different phases.
The dielectric loss is voltage dependant, while the rest is current
dependant.
58 J.R.Lucas
High Voltage Cables
3.1.1 Dielectric loss
For a perfect dielectric, the power factor is zero.
Since the cable is not a perfect dielectric, the power factor is not zero.
The current leads the voltage by an angle of less than 90
o
, and hence there
is a power loss.
If C is the capacitance of the cable, and E is the applied voltage, then

δ

φ

E
I Figure 3.3 -
Loss angle
59 J.R.Lucas
High Voltage Cables
charging current I
c
= E C ω
power loss P = E I cos φ = E I sin δ
= E (I/cos δ) sin δ = E
2
C ω tan δ
The power loss is proportional to E
2
and tan δ .
3.1.2 Conductor loss
The conductor loss P
c
is given by
P
c
= I
2
R
c
watt
where R
c
is resistance of the conductor and I is the current in the cable.
3.1.3 Sheath loss
Losses occurring in the sheath of a cable is usually obtained by the
empirical Arnold's formula P
sh
given by
60 J.R.Lucas
High Voltage Cables
watt
d
r
R
I
10
7.7 =
P
m
2
sh
2
3 -
sh

,
`

.
|
×
where r
m
= mean radius of sheath
d = distance between cables (centre to centre)
R
sh
= resistance of full length of cable
I = current in cable
The sheath loss is usually about 2 to 5 % of the conductor loss.
3.1.4 Intersheath Loss
Intersheath losses are caused by the induced emf between the sheaths
causing a circulating current.
This loss is thus present only when the sheaths of adjacent cables are
connected together.
The sheaths need to be connected together in practice, as otherwise
sparking could occur causing damage to the sheaths.
61 J.R.Lucas
High Voltage Cables
The intersheath loss P
ish
can be calculated in the following manner.
The mutual inductance M
sh
between a core of one cable and the sheath of
an adjacent cable is given by

,
`

.
|

r
d

2
=
Msh
ln
π
µ
The voltage induced E
ish
is given by
E
ish
= I . ω . M
sh
and the induced current I
ish
is given by
[ ] [ ]
M
+
R
M
i
=
M
+
R
E
=
I
sh
2
2
sh
2

sh
sh
2
2
sh
2

ish
ish
2
1
2
1
ω
ω
ω
Therefore the intersheath loss P
ish
is given by
62 J.R.Lucas
High Voltage Cables
R
.
M
+
R
M I
=
R I
=
P sh
sh
2
sh
2
sh
2
2 2
sh ish
2
ish
ω
ω
Generally, the sheath resistance R
sh
>> ω M
sh
so that
R
M I
=
P
sh
sh
2
2 2
ish
ω
The intersheath loss is larger than the sheath loss and may range from
10% to 50% of the copper loss.
Thus the total power loss (exclusive of the dielectric loss) is given as
Total Power loss = P
c
+ P
sh
+ P
ish
R
M I
+
d
r
R
10
x 7.7

I
+ R
I
=
P
sh
sh
2
2 2
m
2
sh
-3
2 2
loss
ω

,
`

.
|
63 J.R.Lucas
High Voltage Cables
Since the whole expression is dependant on I
2
, we may express the loss in
terms of an effective resistance R
eff
.
This gives the total power loss in terms of the effective resistance as
P
total
= I
2
R
eff
R
M
+
d
r
R
10
x 7.7
+
R
=
R
sh
sh
2
2
m
2
sh
-3
c eff
ω

,
`

.
|
Since the sheath loss is usually very small, the
effective conductor resistance can be written
as
3.1.5 Cross-bonding of Cables
R
M
+
R
=
R
sh
sh
2
2
c eff
ω
64 J.R.Lucas
High Voltage Cables
When three single phase cables are used in power transmission, currents are
induced in the sheaths and lead to sheath circulating currents and power loss.
These may be substantially reduced, and the current rating of the cable increased
by cross bonding of the cables (Figure 3.4).
Cross bonding of cables are done except for very short lengths of cable.
The continuity of each cable sheath is broken at regular intervals;
the cables between two adjacent discontinuities is a minor section.

a
Figure 3.4 - Cross
bonding of sheaths bcabcabc
65 J.R.Lucas
High Voltage Cables
Three minor sections make up a major section, where the sheaths are
solidly bonded together and to earth.
A residual sheath voltage exists, and the desired balance, giving negligible
sheath voltage between the solid grounded positions is achieved by
transposing the cables at each cross-bonded section.
To prevent excessive voltage build up at the cross bonded points,
especially during faults, these points are earthed through non-linear
resistors which limit voltage build up. The cable is also transposed
(Figure 3.5)
Figure 3.5 – Non-linear
resistor earthing
66 J.R.Lucas
High Voltage Cables
3.2Impregnated Paper Insulation
The insulation consists mainly of paper tape impregnated with compound.
The paper must be free from ligneous fibres and from metallic or other
conducting spots.
The compound with which the paper is insulated should be of such a
consistency that it is plastic at ordinary temperatures, and has no tendency
to drain away from the cable.
The impregnating compound varies from manufacturer to manufacturer,
but they all are based o paraffinic or naphthenic mineral oil, with resin
frequently added to lower the viscosity and to improve its impregnating
qualities.
The paper is made from Manila fibre or wood pulp.
67 J.R.Lucas
High Voltage Cables
Impregnated paper can withstand an electric stress if about 5 to 10 times
that which could be withstood by dry paper insulation.
The dielectric strength of impregnated paper is about 200 to 300 kV/cm.
Initially, they may be able to withstand about 400 to 600 kV/cm.
The cause of breakdown is usually the non-homogeneity of the dielectric.
When a test voltage is applied, the weakest part of the dielectric
breaksdown and deterioration starts getting more and more.
time
(hrs)
F
i
g
u
r
e

3
.
6


-


B
r
e
a
k
d
o
w
n

v
o
l
t
a
g
e

c
h
a
r
a
c
t
e
r
i
s
t
i
c

o
f

p
a
p
e
r

i
n
s
u
l
a
t
i
o
n
b
.
d
.
v
.
(
k
V
)
5
0
0
w
i
t
h
o
u
t

p
r
e
s
s
u
r
e
4
0
0
3
0
0
2
0
0
1
0
0
0Under
pressure
(15
atmospher
e)
68 J.R.Lucas
High Voltage Cables
This is accentuated by the fact that the cable is not carrying the same
current all the time.
The deterioration results in the formation of voids and gasses.
When the voltage is raised, ionisation or glow discharge can occur in the
voids and ionic bombardment of thee surface.
Some oil suffers condensation and hydrogen and other gases are evolved.
Thus long term breakdown strength and the instantaneous break down
strengths differ.
This value may decrease with time due to deterioration to about 160 to
200 kV/cm.
In the case of a badly impregnated dielectric, the breakdown stress will
continue to decrease and ultimately leads to breakdown.
Using a safety factor, not more than about 40 kV/cm is allowed in service.
F
i
g
u
r
e

3
.
6


-


B
r
e
a
k
d
o
w
n

v
o
l
t
a
g
e

c
h
a
r
a
c
t
e
r
i
s
t
i
c

o
f

p
a
p
e
r

i
n
s
u
l
a
t
i
o
n
69 J.R.Lucas
High Voltage Cables
3.2.1 Properties required of cable insulation
Dielectrics used for cable insulation must have the following properties.
1. High Insulation resistance
2. High dielectric strength
3. Good mechanical strength
4. Immune to attack by acids and alkali in the range 0 - 100
o
C
5. Should not be too costly
6. Should not be hygroscopic (tending to absorb water),
or if hygroscopic should be enclosed in a water tight covering.
3.2.2 Principle underlying the design of high voltage cable insulation
By means of dielectric tests on cables, it has been observed that the long term
breakdown stress is increased if the cable is subjected to pressure.
This is due to the fact that the pressure discourages the formation of voids.
F
i
g
u
r
e

3
.
6


-


B
r
e
a
k
d
o
w
n

v
o
l
t
a
g
e

c
h
a
r
a
c
t
e
r
i
s
t
i
c

o
f

p
a
p
e
r

i
n
s
u
l
a
t
i
o
n
70 J.R.Lucas
tan
δ
0.0
1
02
0
4
0
6
0
8
0
10
0
kV/c
m
applied
stress
(a
)
(b
)
(c
)
100 time
(hrs)
breakdown
stress
(kV/cm)
50
0
40
0
30
0
20
0
10
0
0(a
)
(c
)
(b
)
High Voltage Cables
Even for a badly impregnated cable, the application of pressure improves the
power factor (or loss tangent) considerably.
If the cable is subjected to a pressure of about 15 atmospheres, the long term
dielectric strength improves to about 400 kV/cm and a working stress of about
150 kV/cm may be used (Figure 3.7).
(a) impregnated ( 1 atmos)
(a) 15 atmospheres
(b) badly impregnated ( 1 atmos)
(b) 8 atmospheres
(c) badly impregnated (15 atmos)
(c) 1 atmosphere
Figure 3.7 - Variations with pressure
71 J.R.Lucas
High Voltage Cables
Comparison of curves for (a) well impregnated cable at atmospheric
pressure, (b) badly impregnated cable at atmospheric pressure, and (c)
badly impregnated cable at a pressure of 15 atmospheres for about 47
hours, shows advantages of the pressure on the reduction of power factor.
Further curves show how long term breakdown stress is improved by
pressure.
In modern high voltage cables, using of better materials, power factor has
been reduced from about (0.007 to 0.01) to about (0.002 to 0.003).
For high voltage cables, impregnated paper insulation is very commonly
used.
The paper is porous and contains in itself the impregnating compound.
There are no voids present as the oil is present between the layers of the
paper which forms the insulation.
3.2.3 Paper insulated power cables
72 J.R.Lucas
High Voltage Cables
The conductor of the cable is stranded, and this is lapped round with the
paper tape.
It is first heated to about 100
oC
taking care not to burn it.
A vacuum is then applied for 20 to 50 hours to get rid of any trapped air
inside the cable, and while still under vacuum, impregnating compound is
poured into the tank and thereafter a pressure of 50 p.s.i. (about 0.35
MN/m
2
) is applied.
Impregnating of the paper prevents void formation in the dielectric, as
voids can easily lead to the breakdown of the dielectric.
As paper is hygroscopic, a seamless lead sheath is extruded over the
insulation so that no moisture will get in.
For high voltages, pressurised cables are used where the impregnated
paper insulation is kept under pressure.
73 J.R.Lucas
High Voltage Cables
A pressure of about 15 atmospheres is applied so that any potential voids
would be instantaneously filled. The pressure may be applied by having
either oil or gas under pressure.
When the cable is pressurised, longitudinal reinforcement to prevent
bulging and reinforcement to prevent hoop stress are used.
With pressurised cables, the long term breakdown strength does not differ
much from the short term strength, and as such using a safety factor, a
working stress of about 100 to 120 kV/cm may be used.
74 J.R.Lucas
Δ
x
rRx
High Voltage Cables
3.2.4 Insulation Resistance
F
i
g
u
r
e
3
.
8
-
C
a
b
l
e
c
r
o
s
s
-
s
e
c
ti
o
n
75 J.R.Lucas
High Voltage Cables
For a single core cable (figure 3.8), the insulation resistance between the
conductor and the outer sheath is given by the following.
(km) cable of length = l where
M
10
x )
r
R
( log
l
3.665
= Res . e . i
)
r
R
ln(
l 2
=
x
dx
l 2
= Res = Res
(m) cable of length = l where
10 -
10
R
r



∆ ∫ ∴



ρ
π
ρ
π
ρ
π
ρ

l x 2
x .
= Res
3.2.5 Capacitance in a single-core cable
F
i
g
u
r
e
3
.
8
-
C
a
b
l
e
c
r
o
s
s
-
s
e
c
ti
o
n
F
i
g
u
r
e
3
.
9
-
C
a
b
l
e
c
r
o
s
s
-
s
e
c
ti
o
n
R
76 J.R.Lucas
High Voltage Cables
Consider a single core cable (figure 3.9) with the following data.
r = radius of core (m)
R = radius of earthed sheath (m)
q = charge/unit length of cable (C/m)
D = electric flux density
= charge density (C/m
2
)
ε
0
= permittivity of free space
= 1/(4π x 9 x 10
9
) F/m
Consider an elemental cylinder of radius x and thickness dx, and of length
unity along the cable.
(For impregnated paper insulation, ε
r
= 3.5)
F
i
g
u
r
e
3
.
9
-
C
a
b
l
e
c
r
o
s
s
-
s
e
c
ti
o
n
Δx
r
x
77 J.R.Lucas
High Voltage Cables
( ) ( )
( )
F/km
(R/r) log
0.024
= F/mile
(R/r) log
0.0383
= C
F/m
R/r log
10
x 18
=
V
q
= e capacitnac
R/r log
2
q
= R/r log q
10
x 18
= V . e . i
dx
x
xq
10
x 18
= dx = V that so ,
x d
v d
= also,
x
xq
10
x 18
=
x 2
q
=
D
= stress electric
10
r
10
r
e
9
r
e e
r
9
r
9
R
r
R
r
r
9
r 0
µ
ε
µ
ε
ε
ε π
ε
ε
ξ ξ
ε ε ε
π ε
ξ
π
π




∫ ∫

x 2
q
= D , x1 x 2 x D = q
F
i
g
u
r
e
3
.
9
-
C
a
b
l
e
c
r
o
s
s
-
s
e
c
ti
o
n
78 J.R.Lucas
High Voltage Cables
Field Plotting
In the case of the single core cables, the stress is radial, and its magnitude
alternates with time.
The electric flux lines and the equipotential lines are perpendicular to each
other.
Further for constant differences, ∆ φ (or q) and ∆ V are constants.
Thus for the elemental figure shown, capacitance is constant.
x
l y
d
A
C


· ·
ε ε .
= constant.
equipot
ential
lines
electric
flux
lines

y

x

φ

V
79 J.R.Lucas
High Voltage Cables
i.e. y/∆ x = constant (usually chosen as 1 for convenience of drawing).
Thus curvilinear squares are formed in the sketch.
In the case of the 3-core cable, since the centres of the cores lie in a circle,
the electrostatic field is a somewhat rotating field and not a pulsating one.
Typical variations of the equipotential surfaces, for a few points of the
cycle are illustrated in figure 3.10.
From these it will be seen that the field lines, which are perpendicular to
the equipotential lines, are not radial to the individual cores.
Consequently, the electric stress is not radial, and tangential components
of stress exist.
If paper insulation is used around each cores, then tangential stresses will
be applied along the surface of the paper rather than just across it.
The electrical properties of paper varies in different directions.
80 J.R.Lucas
High Voltage Cables
The effective dielectric stress of paper insulation is much greater across
the layers than along it.
Thus the presence of tangential stress in paper insulation leads to greater
risk of breakdown.
81 J.R.Lucas
High Voltage Cables
3.2.6 Three-core Cables
When three phase power is being transmitted, either three single-core cables or a
single three-core cable may be used.
82 J.R.Lucas
High Voltage Cables
ABC
equipotential lines electric stress lines
1.
0
-
0.5
-
0.5
83 J.R.Lucas
High Voltage Cables
84 J.R.Lucas
1∠
30
o
123.123
High Voltage Cables
V
1
= 0.866
V
2
= 0
V
3
= – 0.866
85 J.R.Lucas
High Voltage Cables
Figure 3.10 - Equipotential lines in three-core cables
1∠ 1
5
o
123.123
V
1
= 0.966
V
2
= – 0.259
V
3
= – 0.707
86 J.R.Lucas
High Voltage Cables
3.2.7 Three-core belted type Cables
In the case of a 3-core cable, the 3-cores are individually insulated with
paper insulation.
The filler spaces between the core insulation is also filled up with
insulation, but depriving these of voids is much more difficult.
Belt insulation is used on top of all three core insulations, and the lead
sheath is extruded over this.
Over the lead sheath, there is generally bitumen to prevent damage.
In buried cables, additional protection is necessary to prevent damage.
There are two types of armouring used for these cables.
(i) Steel tape armouring - steel tape is usually wound in two
layers with opposite directions of lay
(ii) Steel wire armouring - steel wires are laid in one or two layers.
87 J.R.Lucas
High Voltage Cables
Capacitance of 3-core belted type
c
o
n
d
u
c
t
o
r
S
h
e
a
t
h
F
i
g
u
r
e
3
.
1
1
-
3
c
o
r
e
b
e
lt
e
d
c
a
b
l
e
c
o
r
e
i
n
s
u
l
a
ti
o
n
b
e
lt
i
n
s
u
l
a
ti
o
n
88 J.R.Lucas
High Voltage Cables
The capacitance between the conductor to neutral of 3-core belted cables
(Figure 3.11) cannot be obtained by a simple derivation as for the single
core cable. Simon's expression can be used to obtain this value.
The capacitance per unit length to neutral is given by
If t = thickness of belt insulation
T = thickness of conductor insulation
d = diameter of conductor
ε
r
= dielectric constant
F/km
log
µ
ε

1 +
d
t + T
3.84 +
T
t
1.7 -
T
t
0.52
0.03
=
C
2
10
r
0
]
]
]
]

,
`

.
|

,
`

.
|

,
`

.
|

,
`

.
|
Measurement of capacitance of 3-core cables
F
i
g
u
r
e
3
.
1
1
-
3
c
o
r
e
b
e
lt
e
d
c
a
b
l
e
89 J.R.Lucas
High Voltage Cables
Figure 3.12 - Cable Capacitances C
s
CC
C
s
C
s
C
90 J.R.Lucas
High Voltage Cables
In three-core cables, capacitance does not have a single value, but can be
lumped as shown in figure 3.12.
C between each core and sheath = C
s
C between cores = C
These Capacitances can be separated from measurements as described in
the following section.
(a) Strap the 3 cores together and measure the capacitance between this
91 J.R.Lucas
High Voltage Cables
(a) Strap the 3 cores together and measure the capacitance between this
92 J.R.Lucas
High Voltage Cables
C
s
C
s
C
s Bri
dge
Figure 3.13 -
Capacitance
measurement
93 J.R.Lucas
High Voltage Cables
(a) Strap the 3 cores together and measure the capacitance between this
bundle and the sheath as shown in figure 3.13.
Measured value = C
m1
= 3 C
s
Gives capacitance to the sheath as
C
s
= C
m1
/3
(b) Connect 2 of the cores to the sheath and measure between the
94 J.R.Lucas
High Voltage Cables
(b) Connect 2 of the cores to the sheath and measure between the
95 J.R.Lucas
High Voltage Cables
Bri
dge
Figure 3.14 -
Capacitance
measurement
C
s
CCta
n
δ
96 J.R.Lucas
High Voltage Cables
(b) Connect 2 of the cores to the sheath and measure between the
remaining core and the sheath (Figure 3.14).
Measured value C
m2
= 2 C + C
s
i.e. C = (C
m2
- C
s
)/2 = (3 C
m2
- C
m1
)/6
giving the capacitance between the conductors.
Effective capacitance to neutral C
o
of any cores may be obtained by
97 J.R.Lucas
High Voltage Cables
Effective capacitance to neutral C
o
of any cores may be obtained by
98 J.R.Lucas
C
s
C
s
C
s
3
C
3
C
3
C
High Voltage Cables
Figure 3.15 -
Calculation of C
o
99 J.R.Lucas
High Voltage Cables
Effective capacitance to neutral C
o
of any cores may be obtained by
considering the star equivalent (Figure 3.15).
1
C
6
1
- 2
C
2
3
=
C
6
1
C
- 2
C
3
3 + 1
C
3
1
= C 3 +
C
=
C
m m 0
m m
m s 0
In the breakdown of actual 3-core belted cables, it is generally observed
that charring occurs at those places where the stress is tangential to the
layers of paper.
Thus for the insulation to be effective, the tangential stresses in paper
insulation should be preferably avoided.
Can usually be accomplished by screening each core separately (or by
having individual lead sheaths for each core), so that the cable in effect
becomes 3 individual cables laid within the same protective covering.
100 J.R.Lucas
High Voltage Cables
3.2.8 Hochstadter or "H" type Cable
In this type of cable (Figure 3.16), there is no belt insulation.
The screening of individual cores is generally thin and flexible so that
there is not much power dissipation in them.
Figure 3.16 -
H-type cable
paper
worming
copper
woven
fabric
paper
insulatio
n
metallise
d paper
lead
sheath
conduct
or
101 J.R.Lucas
High Voltage Cables
All the individual screens are earthed so that the potential at these sheaths
are all zero and thus the stress lines between the cores and screens would
be now radial.
The screens are thin so that there is hardly any current induced. The
sheaths surrounding the insulation of the cores consist of metallised
perforated paper.
These are wrapped round with copper woven fabric (cotton tape into
which are woven copper wire).
This outer screen is in contact with the inner screens and is earthed.
The cable has the additional advantage that the separation of the cores by
thermal expansion or mechanical displacement cannot introduce stresses
in the dielectric.
The metallised screens help to dissipate the heat.
These are used upto 66 kV.
In the H-type cable, the individual cores contain no lead covering.
102 J.R.Lucas
High Voltage Cables
The three cores are laid up with fillers in the ordinary way.
If cable is to be buried, then cable is armoured with steel wire and tape.
The wormings of the H-type cable are full of oil.
3.2.9 S.L. type Cable
Figure 3.17 – S
L type cable
compou
nd jute
worming
core
insulatio
n
cotton
tape
lead
sheath
cor
e
103 J.R.Lucas
High Voltage Cables
Another development of the screening principle is the SL type cable
(Figure 3.17).
In this, each core is screened and then individually sheathed with lead or
aluminium.
These do not have an overall lead sheath.
The electric field in the insulation surrounding each core is naturally radial and
the function of the screens in this case is to eliminate the possibility of any stress
across the clearance space between core and sheath.
The wormings of the filler spaces in the S.L. type cables do not contain much oil
as do not get any electric stress.
The three metal sheathed cores, after being lapped with paper and cotton
tapes are laid with tarred jute yarn to get a circular formation and then
wrapped with hessian tapes to form a bedding for the armouring.
The electrical and thermal advantages of H-type cables are also enjoyed
by the S.L. type cables.
104 J.R.Lucas
High Voltage Cables
These cables are suitable for hilly routes, as the absence of oil in the filler
spaces lessens the risk of oil drainage.
Also, the S.L. type construction is useful on short runs because the
terminating equipment is simplified.
Also the void formation in the filler spaces are of no consequence.
The separate lead sheaths in the S.L. cable are the seats of induced
currents, but the resulting losses are small, and appear to be of no practical
significance.
105 J.R.Lucas
High Voltage Cables
3.2.10 Copper Space Factor
Unlike in overhead lines, insulation in cables occupies a greater portion of
the cable space.
Thus higher installation costs are involved.
F
i
g
u
r
e
3
.
1
8
rR
106 J.R.Lucas
RrTt
High Voltage Cables
Ideally we would like the insulation to occupy the minimum possible
thickness.
Thus we define a space factor to indicate the utilisation of the space.
copper space factor =
cable whole of area tion cross
conductor of area tion cross
sec
sec


For a single core cable, the best space factor is obtained with a concentric
arrangement (Figure 3.18), as this gives the minimum conductor perimeter for
the greatest conductor area and given insulation thickness.
Thus Space factor = r
2
/R
2
F
i
g
u
r
e
3
.
1
9
-
t
h
r
e
e
-
c
o
r
e
c
a
b
l
e
107 J.R.Lucas
High Voltage Cables
For the 3-core cable (Figure 3.19), consisting of circular conductors
within a circular sheath,
Space factor = 3 r
2
/R
2
where T = thickness of core insulation,
t = thickness of belt insulation
and R
1
= r + T
For the 3-core cable circular cross-section is not the best shape for the
conductors.
Other shapes which gives better space factors are the elliptical shaped
conductors and the sector shaped conductors (Figure 3.20).
F
i
g
u
r
e
3
.
1
9
-
t
h
r
e
e
-
c
o
r
e
c
a
b
l
e
F
i
g
u
r
e
3
.
2
0
-
S
p
e
c
i
a
l
s
h
a
p
e
s
o
f
c
o
n
d
u
c
t
o
r
s
t
o
g
i
v
e
b
e
tt
e
r
s
p
a
c
e
f
a
c
t
o
r
s
108 J.R.Lucas
High Voltage Cables
F
i
g
u
r
e
3
.
1
9
-
t
h
r
e
e
-
c
o
r
e
c
a
b
l
e
F
i
g
u
r
e
3
.
2
0
-
S
p
e
c
i
a
l
s
h
a
p
e
s
o
f
c
o
n
d
u
c
t
o
r
s
t
o
g
i
v
e
b
e
tt
e
r
s
p
a
c
e
f
a
c
t
o
r
s
109 J.R.Lucas
High Voltage Cables
3.3Dielectric Stress in a Single Core Cable
The voltage difference across the conductor and the sheath of a single core
cable is given by
r
R
log x
v
= that so
x 2
q
= , also log
e
x
x
ξ
ε π
ξ
ε π
,
r
R

2
q
= v
e
It is seen that since x is the only variable, the maximum stress in the
dielectric occurs at the minimum value of the radius x (i.e. x = r).
Since it is required that this maximum stress in the dielectric should be as
low as possible, differentiating with respect to r for minimum ξ
max
gives
F
i
g
u
r
e
3
.
2
0
-
S
p
e
c
i
a
l
s
h
a
p
e
s
o
f
c
o
n
d
u
c
t
o
r
s
t
o
g
i
v
e
b
e
tt
e
r
s
p
a
c
e
f
a
c
t
o
r
s
110 J.R.Lucas
High Voltage Cables
2.718 = e =
r
R
. e . i
0 =
r
1
- . r +
r
R
log .
R/r log r
V
. e . i
e
e
2
max
]
]
]

,
`

.
|
]
]
]

0 =
r d
d ξ
15
Thus if the overall diameter of the cable is kept fixed, then R/r = e is the
condition for minimum ξ
max
.
This value of radius of conductor will generally be larger than would be
required for current carrying capacity.
Since R/r = e, the minimum value of ξ
max
is given by
F
i
g
u
r
e
3
.
2
0
-
S
p
e
c
i
a
l
s
h
a
p
e
s
o
f
c
o
n
d
u
c
t
o
r
s
t
o
g
i
v
e
b
e
tt
e
r
s
p
a
c
e
f
a
c
t
o
r
s
111 J.R.Lucas
High Voltage Cables
r
V
=
R/r r
V
=
e
log
max
ξ
F
i
g
u
r
e
3
.
2
0
-
S
p
e
c
i
a
l
s
h
a
p
e
s
o
f
c
o
n
d
u
c
t
o
r
s
t
o
g
i
v
e
b
e
tt
e
r
s
p
a
c
e
f
a
c
t
o
r
s
112 J.R.Lucas
ξξ
m
High Voltage Cables
R x r0
F
i
g
u
r
e
3
.
2
1
-
S
tr
e
s
s
D
is
tr
i
b
u
ti
o
n
113 J.R.Lucas
High Voltage Cables
Since the radius of the conductor that would be given from the above
expression is larger than is necessary for current carrying capacity, this
value of radius may be achieved by using Aluminium or hollow
conductors.
As can be seen (Figure 3.21), the dielectric is not equally stressed at all
radii, in a cable of homogeneous insulation. The insulation is fully
stressed only at the conductor, and further away near the sheath the
insulation is unnecessarily strong and thus needlessly expensive.
F
i
g
u
r
e
3
.
2
1
-
S
tr
e
s
s
D
is
tr
i
b
u
ti
o
n
114 J.R.Lucas
High Voltage Cables
3.3.1 Cable Grading for Uniform Stress Distribution
The electric stress in the dielectric may be more equally distributed by one
of the two following methods.
(i) Capacitance grading
(ii) Intersheath grading
3.3.2 Capacitance Grading
In this method of grading, the insulation material consists of various
layers having different permittivities.
Consider a cable graded by means of 3 layers of insulation, as shown in
Figure 3.22, having permittivities ε
1
, ε
2
, ε
3
, respectively.
Let the outer radii of these layers by r
1
, r
2
and r
3
= R respectively,
and the conductor radius r.
115 J.R.Lucas
ξ
ξ
m1
ξ
m2
ξ
m3
R
x
r r
1

r
2
0
High Voltage Cables
18
F
i
g
u
r
e
3
.
2
2
C
a
p
a
c
it
a
n
c
e
G
r
a
d
i
n
g
116 J.R.Lucas
High Voltage Cables
In order to secure the same value of maximum stress in each layer, the
maximum stresses in the layers are equated.
r
=
r
= r

r
2
q
=
r
2
q
=
r 2
q
2 3 1 2 1
2 3 0 1 2 0 1 0
ε ε ε
ε ε
π
ε ε
π
ε ε
π

Let the voltage across the inner-most layer of insulation be V
1
. Then
z
determined be can
V
,
V
similarly
log
3 2
max
r
r
r =
V
1
e
1
ξ
F
i
g
u
r
e
3
.
2
2
C
a
p
a
c
it
a
n
c
e
G
r
a
d
i
n
g
117 J.R.Lucas
High Voltage Cables
Therefore the total voltage across the dielectric can be obtained as
follows.
r >
r
r, >
r
since log
ln ln ln
ln ln ln ln ln
ln ln ln
2 1
max
max
max
max
,
r
R
r >
r
R
r) -
r
( +
r
r
r) -
r
( +
r
R
r =
r
R
r) -
r
( +
r
r
r) -
r
( +
r
R
r +
r
r
r +
r
r
r =
r
r
r
+
r
r
r
+
r
r
r = V
e
2
2
1
2
1
2
2
1
2
1
2 1
2 1
2
3
2
1
2
1
1
ξ
ξ
ξ
ξ

,
`

.
|

,
`

.
|

,
`

.
|
Hence by grading the insulation, without increasing the overall diameter of the
cable, the operating voltage can be raised.
A difficulty with this method is that we cannot obtain a wide range of
permittivities in practice, as paper insulation has permittivities limited to the
range 2.8 to 4.0.
118 J.R.Lucas
High Voltage Cables
In the above analysis, it has been assumed that the maximum permissible
stress is the same for all three dielectrics used.
If the maximum stress in the three sections are different, and are ξ
1
, ξ
2
, ξ
3
respectively, then the maximum stresses should be reached at the same
time for the most economical operation of the insulation.
This condition gives us the result
r
=
r
= r
2 3
3
1 2
2
1
1
ε
ξ
ε
ξ
ε
ξ
119 J.R.Lucas
R
x
ξξ
m
r r
1

r
2
0V
3
V
2
V
1
rr
1
r
2
RV
High Voltage Cables
3.3.3 Intersheath Grading
F
i
g
u
r
e
3
.
2
3

I
n
te
rs
h
e
at
h
G
r
a
d
i
n
g
120 J.R.Lucas
High Voltage Cables
In this method of grading, the same insulating material is used throughout
the cable, but is divided into two or more layers by means of cylindrical
screens or intersheaths (Figure 3.23).
These intersheaths are connected to tappings from the supply transformer,
and the potentials are maintained at such values that each layer of
insulation takes its proper share of the total voltage.
The intersheaths are relatively flimsy, and are meant to carry only the
charging current.
Since there is a definite potential difference between the inner and outer
radii of each sheath, we can treat each section separately as a single core
cable.
If V
1
, V
2
, V
3
, .... are the potential differences across the sections of
insulation, then
F
i
g
u
r
e
3
.
2
3

I
n
te
rs
h
e
at
h
G
r
a
d
i
n
g
121 J.R.Lucas
High Voltage Cables
..... =
r
r
r
V
=
r
r
r
V
=
1
2
e
2
1
e
1
log log
max
ξ
Since the cable insulation now consists of a number of capacitors in
series, formed by the respective intersheaths, all potential differences V
1
,
V
2
, V
3
, ... are in phase.
Thus, if V is the phase to neutral voltage, we can also write
V = V
1
+ V
2
+ V
3
+ . . . . . . . . + V
n
In the particular case that all the n layers have the same thickness d, and if
r is the conductor radius,
r
1
= r + d; r
2
= r + 2d; r
3
= r + 3d; ....... r
n
= r + n d
122 J.R.Lucas
High Voltage Cables
( )
d 1) - (m + r
d m + r
log d 1) - (m + r = M where
log log
e
n
1 = m
max

M
V
= ..... =
d + r
d 2 + r
d) + (r
V
=
r
d + r
r
V
=
e
2
e
1
ξ
The voltage across the m
th
section is given by
[ ]
d 1) - (m + r
d m + r
d 1) - (m + r
M
V
=
V
e
m
log
Hence substituting for the different values of m, we can obtain the voltage
across the various layers that have to be maintained to give equal
maximum stress in each section.
In practice, there is a considerable difficulty in arranging for many
intersheaths, this difficulty being mainly associated with the provision of
123 J.R.Lucas
High Voltage Cables
the different voltages for the intersheaths, and as a result it is usual to
design a cable of this type with only one intersheath.
This simplifies the design calculations, and the expression for the
maximum stress then given by
r
r
r +
r
R

r
V
=
1
e
1
e
1
log log
max
ξ
For the purpose of comparison with the ungraded cable, let us first take
the optimally designed ungraded cable (i.e. with R/r = e), and introduce
an intersheath at a radius r
1
.
Since R and r are both kept fixed, r
1
is the only variable, and the
expression for stress must be differentiated with respect to r
1
to obtain the
condition for the minimum value of the maximum stress.
124 J.R.Lucas
High Voltage Cables
1 =
r
r
log
r
r

1 = e log also e, =
r
R
g considerin
0 =
r
r
+
r
R
log + 1 - . e . i
1
e
1
e
1 1
e

0 =
r
1
.
r
r
. r +
r
R
log +
r
R -
.
R
r
.
r
that so
1 1
e
1
2
1
1
max

,
`

.
|


0 =
r1
ξ
24
This gives the solution r
1
= 1.76 r.
r 1.33
V
=
r
r 1.76
r +
r 1.76
r e
r 1.76
V
=
e e

,
`

.
|

,
`

.
|

log log
max
ξ
125 J.R.Lucas
High Voltage Cables
25
However, for the cable without intersheath, we have ξ'
max
= V/r.
Hence, the addition of the intersheath raises the maximum applicable
voltage by 33%.
Consider the case of only overall diameter R being fixed, and both r and
r
1
being variable. Then for minimum value of maximum stress we have
1.881 =
r
R
giving
e
1
- 1 =
r
R
log gives This
0 =
e
1
+
r
R
log + 1 - e, =
r
r
. e . i
0 =
r
r
+
r
R
log + 1 - also 0, = 1 -
r
r
log . e . i
1 1
e
1
e
1
1 1
e
1
e
max max
0 =
r
0, =
r
1



∂ ξ ξ
126 J.R.Lucas
High Voltage Cables
r 2.718
V
= . e . i
log log
max
max
ξ
ξ

r e
V
=
r +
e
1
- 1 r e
V
=
e r +
r
R
r e
V
=
e
1
e

,
`

.
|
127 J.R.Lucas
High Voltage Cables
27
3.4Pressurised High Voltage Cables
In high voltage paper insulated cables, the application of pressure (about
13 atmospheres) increases the maximum allowable working stress (after
applying a suitable safety factor) from about 50 kV/cm to about 150
kV/cm.
In super voltage cables, the void control is effected by pressurising the oil-
impregnated paper tape insulation by
(a) pressurising the oil, and
(b) applying gas pressure.
128 J.R.Lucas
High Voltage Cables
3.4.1 Oil-pressure cables
In oil filled cables, the oil must be free to flow inorder to transmit the
pressure.
The maximum pressure of oil utilised is about 0.35 MN/m
2
(3.5
atmospheres or 50 p.s.i.).
Due to the pressure of oil, the sheath tends to bulge out and therefore
reinforcement of the sheathing is necessary.
A reservoir maintains the required pressure.
The cable can now operate at a maximum working stress of 150 kV/cm.
In normal, solid type of cable, the drying and impregnating are done
before sheathing, while in oil-filled cables they can be done after
sheathing by circulating hot oil.
The oil filled construction permits a great reduction in size of the cable.
129 J.R.Lucas
High Voltage Cables
There are 3 main types of oil filled cables.
These are (a) single-core, conductor channel; (b) single-core, sheath
channel; and (c) three-core, filler-space channels.
(a) Single-core conductor channel
This type of cable shown in figure 3.24 has a hollow conductor which acts
as an oil channel, and is the simplest from the point of view of the cable
itself. A disadvantage of this arrangement is that the oil is at high voltage
oil duct hollow
conductor
paper
insulation
lead
sheath
helical
ribbon
reinforcement Figure 3.24 – Single core conductor channel cable
130 J.R.Lucas
High Voltage Cables
with respect to earth being at the voltage of the conductor. The copper
strands of which the conductor is made are laid over a helical metal ribber,
so that oil can reach the insulation.
(b) Single-core sheath channel
131 J.R.Lucas
High Voltage Cables

grooved
lead
sheath
oil
channels
lead
sheath
p
a
p
e
r
i
n
s
u
l
a
ti
o
n
button-stamped
helical spacer ribbon
F
i
g
u
r
e
3
.
2
5
-
S
i
n
g
l
e
c
o
r
e
s
h
e
a
t
h
c
h
a
n
n
e
l
c
a
b
l
e
132 J.R.Lucas
High Voltage Cables
In this type (Figure 3.25), the oil channels are produced either by grooving
the sheath or by arranging spacers between sheath and insulation.
The resistance to oil flow in this type is 6 to 8 times that of type (a), so
that more feeding points are necessary to maintain the pressure.
An advantage is that the channels are at earth potential so that joints and
installation are simpler.
(c) Three-core, filler space channels
F
i
g
u
r
e
3
.
2
5
-
S
i
n
g
l
e
c
o
r
e
s
h
e
a
t
h
c
h
a
n
n
e
l
c
a
b
l
e
133 J.R.Lucas
High Voltage Cables

F
i
g
u
r
e
3
.
2
5
-
S
i
n
g
l
e
c
o
r
e
s
h
e
a
t
h
c
h
a
n
n
e
l
c
a
b
l
e
l
e
a
d
s
h
e
a
t
h
p
a
p
e
r
i
n
s
u
l
a
ti
o
n
p
a
p
e
r
f
il
l
e
r
p
e
r
f
o
r
a
t
e
d
o
il
-
d
u
c
t
s
F
i
g
u
r
e
3
.
2
6
-
T
h
r
e
e
c
o
r
e
fi
ll
e
r
s
p
a
c
e
c
h
a
n
n
e
l
c
a
b
l
e
134 J.R.Lucas
High Voltage Cables
In this type (Figure 3.26), the oil channels are located in the filler spacers.
These channels are composed of perforated metal-ribbon tubing and are at
earth potential.
p
e
r
f
o
r
a
t
e
d
o
il
-
d
u
c
t
s
F
i
g
u
r
e
3
.
2
6
-
T
h
r
e
e
c
o
r
e
fi
ll
e
r
s
p
a
c
e
c
h
a
n
n
e
l
c
a
b
l
e
135 J.R.Lucas
High Voltage Cables
3.4.2 Gas-pressure cables
In Gas pressure cables, a pressure of about 1.4 MN/m
2
(14 atmospheres or
200 p.s.i.) is used.
Figure 3.27 shows the different types of gas pressure cables.
F
i
g
u
r
e
3
.
2
6
-
T
h
r
e
e
c
o
r
e
fi
ll
e
r
s
p
a
c
e
c
h
a
n
n
e
l
c
a
b
l
e
Gas Pressure (14
atmospheres)
External
Pressure
Internal
Pressure
Pipe line
type
Self-
contained
type
High
pressure
gas-filled
Impregna
ted
pressurise
d
Figure 3.27 - Types of
gas pressure cables
Gas
cush
ion
136 J.R.Lucas
High Voltage Cables
3.4.3 External Pressure Cables
Pipe line type
137 J.R.Lucas
High Voltage Cables
F
i
g
u
r
e
3
.
2
8
-
P
i
p
e
li
n
e
t
y
p
e
c
a
b
l
e
triangular lead
sheath (membrane)
t
h
i
n
m
e
t
a
l
t
a
p
e
Nitrogen at
200 p.s.i.
S
t
e
e
l
p
i
p
e
-
li
n
e
138 J.R.Lucas
High Voltage Cables
The cable, shown in Figure 3.28, is manufactured in the usual way and the
outside is made triangular, and covered by a diaphragm lead sheath.
The pipe is filled with Nitrogen subjected to a pressure of 200 p.s.i. which
is transmitted to the insulation through the diaphragm.
F
i
g
u
r
e
3
.
2
8
-
P
i
p
e
li
n
e
t
y
p
e
c
a
b
l
e
139 J.R.Lucas
High Voltage Cables
The steel pipe is laid first, and the cable is drawn in afterwards.
Nitrogen under pressure is then introduced into the pipe.
The pressure is transmitted to the membrane through the membrane.
In the Self-contained type, an additional reinforced lead sheath is used,
but otherwise the principle is the same as that of the pipe line type.
140 J.R.Lucas
High Voltage Cables
3.4.4 Internal Pressure Cables
In the internal pressure cables, the gas is in contact with the dielectric.
(a) Gas filled cables
141 J.R.Lucas
High Voltage Cables

F
i
g
u
r
e
3
.
2
9
-
G
a
s
fi
ll
e
d
c
a
b
l
e
c
o
tt
o
n
t
a
p
e
r
u
b
b
e
r
t
a
p
e



reinforcement
tape
Paper
impregnation
copper woven
fabric
A
n
n
u
l
a
r
g
a
s
p
a
s
s
a
g
e
lead alloy
sheath
142 J.R.Lucas
High Voltage Cables
In these cables (Figure 3.29), spaces are left between the convolutions so
that the gas is between them.
The presence of Nitrogen prevents the formation of voids.
The method of manufacture is such that the gas can move freely inside
packets, but cannot diffuse outside the insulation.
(b) Gas cushion type
F
i
g
u
r
e
3
.
2
9
-
G
a
s
fi
ll
e
d
c
a
b
l
e
r
u
b
b
e
r
t
a
p
e



A
n
n
u
l
a
r
g
a
s
p
a
s
s
a
g
e
143 J.R.Lucas
High Voltage Cables

s
e
p
a
r
a
t
e
g
a
s
c
u
s
h
i
o
n
s
s
e
a
l
e
d
F
i
g
u
r
e
3
.
3
0
-
G
a
s
C
u
s
h
i
o
n
T
y
p
e
c
a
b
l
e
144 J.R.Lucas
High Voltage Cables
(Type shown is not of much practical use but only of academic use).
In this type, a screened space is provided between the lead sheath and the
dielectric, this space providing accommodation at all points along the
length of the cable for the storage of inert gas under pressure.
This storage is maintained by the subdivision of the screened space into a
series of gas cushions by means of barriers, with the result that the cable
may be cut for joining without losing gas from more than a short length.
(c) Impregnated pressurised cable
s
e
p
a
r
a
t
e
g
a
s
c
u
s
h
i
o
n
s
F
i
g
u
r
e
3
.
3
0
-
G
a
s
C
u
s
h
i
o
n
T
y
p
e
c
a
b
l
e
145 J.R.Lucas
High Voltage Cables

F
i
g
u
r
e
3
.
3
1
-
I
m
p
r
e
g
n
a
t
e
d
P
r
e
s
s
u
ri
z
e
d
c
a
b
l
e
g
a
s
c
h
a
n
n
e
l
l
e
a
d
s
h
e
a
t
h
m
e
t
a
l
s
c
r
e
e
n
t
a
p
e
b
i
n
d
e
r



C
o
p
p
e
r
w
o
v
e
n
f
a
b
r
i
c
(
ti
n
n
e
d
c
o
p
p
e
r
)
s
t
r
a
n
d
e
d
c
o
n
d
u
c
t
o
r
m
e
t
a
ll
i
s
e
d
p
a
p
e
r
s
c
r
e
e
n
146 J.R.Lucas
High Voltage Cables
In the manufacture of this type of cable (Figure 3.31), provision is made
for longitudinal gas flow.
The impregnating compounds used are suitable for the higher dielectric
stresses necessary for high voltage cables.
The cable has a mass impregnated paper dielectric and the impregnating
oil is maintained under a pressure of 200 p.s.i. by means of nitrogen.
Special reinforcement is provided to cater for the large hoop and
longitudinal stresses set up.
The core is stranded and is covered with a metallised paper screen so as to
obtain a completely uniform stress.
The gas channel is in one of the filler spaces.
F
i
g
u
r
e
3
.
3
1
-
I
m
p
r
e
g
n
a
t
e
d
P
r
e
s
s
u
ri
z
e
d
c
a
b
l
e
t
a
p
e
b
i
n
d
e
r



C
o
p
p
e
r
w
o
v
e
n
f
a
b
r
i
c
s
t
r
a
n
d
e
d
c
o
n
d
u
c
t
o
r
m
e
t
a
ll
i
s
e
d
p
a
p
e
r
s
c
r
e
e
n
147 J.R.Lucas
High Voltage Cables
3.5Thermal Design of Cables
Underground cables are installed in trenches of rectangular cross-section.
After excavation of the trench, a layer of sand is placed in it to serve as a
bedding, as shown in Figure 3.32.
F
i
g
u
r
e
3
.
3
1
-
I
m
p
r
e
g
n
a
t
e
d
P
r
e
s
s
u
ri
z
e
d
c
a
b
l
e
Figure 3.32 - Cross-section of Trench and buried cable
Trench width sheat
h
conducto
r
insulation sand
bedding
cover tile depth of
burial
back fill Air (T
A
) servin
g
undisturbed
ground
148 J.R.Lucas
High Voltage Cables
The length of cable is pulled in along the trench and covered with a
further layer of sand.
Sand free from flints and stone is employed to avoid damage to the cable
serving during pulling and initial back filling.
Above the cable and sand bedding are placed cover tiles to protect the
cable from mechanical damage from subsequent excavation activities.
The excavated material is replaced in the trench and stamped to
consolidate it.
The minimum trench width that can be conveniently excavated is about
700 mm (27 inches), and for safety reasons, the minimum depth of burial
in normal circumstances is 900 mm (36 inches).
An underground cable carrying current will have in addition to the
conductor loss, dielectric loss and losses in the sheath.
These produce heat which are conducted away from the cable to the
surface, producing a temperature gradient.
149 J.R.Lucas
High Voltage Cables
When more than one single core cable is laid together (as is required for
three phase systems exceeding 150 kV), the heat produced by one
conductor affects the other and the heat factors need to be modified.
When the spacing between the cables is increased, the heat produced by
the circulating currents between the cables will increase whereas the eddy
current losses decrease.
Thus there is an optimum spacing for cables and various alternatives may
have to be evaluated before the economic arrangement is finally selected.
150 J.R.Lucas
High Voltage Cables
3.5.1 Current rating of Cables
In a cable, the factor which ultimately limits the current carrying capacity is the
maximum operating temperature which may be sustained by the cable
throughout its life without risk of damage or deterioration.
As was discussed in an earlier section, the heat generated in the cable is due to
(a) ohmic loss in the conductor,
(b) the dielectric loss in the insulating medium and
(c) the sheath and intersheath losses.
The heat so generated is radiated to the surroundings.
The current that can be carried depends on the conductivity of the
surrounding medium as well, so that the same cable would have different
ratings depending on whether the cable is buried or not.
3.5.2 Thermal Resistance
151 J.R.Lucas
High Voltage Cables
Since the flow of heat can be considered analogous to the flow of charge
or current in the insulation, the thermal resistance of the cable and
surroundings is measured in terms of the thermal ohm.
Thermal Ohm: The thermal ohm is the resistance of a path through
which a temperature difference of 1
0
C produces a heat flow of 1 watt.
Thermal Resistivity: The thermal resistivity is the temperature drop in
degree centigrade produced by the flow of 1 watt between the opposite
faces of a metre cube of the material.
Consider a cable buried under the surface of the earth.
θ = maximum allowable temperature difference between core and
surroundings (
o
C)
R
θ
= Effective Resistance of conductor (including effects of sheath loss)
I = Current carried by conductor
H = Heat produced in the core (W)
152 J.R.Lucas
High Voltage Cables
S' = Thermal resistance of dielectric
S" = Thermal resistance of cable outside dielectric
S = S' + S" = Total thermal resistance of cable
G = Thermal resistance of ground from cable to surroundings
153 J.R.Lucas
High Voltage Cables

F
i
g
u
r
e
3
.
3
3
-
H
e
a
t
fl
o
w

li
n
e
s
fr
o
m

b
u
ri
e
d
c
a
b
l
e
θ
154 J.R.Lucas
High Voltage Cables
From the definition, the total temperature rise between the conductor and
the surroundings is given by
θ = H (S + G)
Total power loss = dielectric loss (W
d
) + ohmic loss (I
2
R
θ
)
At equilibrium, the total power loss must equal to the heat produced.
G) + (S )
R I
+
W
( = . e . i

2
d θ
θ
θ
θ
R I
+
W
=
G + S
= H

2
d

This gives the current rating of the cable as
G) + (S
R
G) + (S
W
-
= I

d
θ
θ
F
i
g
u
r
e
3
.
3
3
-
H
e
a
t
fl
o
w

li
n
e
s
fr
o
m

b
u
ri
e
d
c
a
b
l
e
155 J.R.Lucas
High Voltage Cables
In calculating the flow if heat it is useful to to remember the following
analogies.
Heat Electricity Electrostatic Electromagnetic
Temperature
Difference θ
Potential
difference
Potential
difference
magnetomotive
force (mmf)
Heat Flow H Current I
Electric
charge Q
Magnetic flux φ
Thermal
resistance S
Resistance 1/Capacitance Reluctance
If a method exists to study any of the above phenomena, the analogous
quantity can also be studied by comparison.
F
i
g
u
r
e
3
.
3
3
-
H
e
a
t
fl
o
w

li
n
e
s
fr
o
m

b
u
ri
e
d
c
a
b
l
e
156 J.R.Lucas
High Voltage Cables
3.5.3 Thermal Resistance of single-core cable
157 J.R.Lucas
r
1
r
2
xd
x
Hr
1
r
2
xd
x
q
High Voltage Cables

equi-temperature
lines
equi-potential
lines
lines of
heat flow

lines of
charge flow

F
i
g
u
r
e
3
.
3
4
-
A
n
a
l
o
g
o
u
s
h
e
a
t-
fl
o
w

a
n
d
c
h
a
r
g
e
-
fl
o
w

li
n
e
s
158 J.R.Lucas
High Voltage Cables
The analysis of this problem is similar to the analysis of the analogous
electrostatic case.
Figure 3.34 shows these two cases.
In the heat problem, H is the amount of heat generated per unit length of
cable, and in the corresponding electrostatic case, q is the electric flux
flowing out per unit length.
For the electrostatic case, consider a gaussian cylinder of radius x and
thickness dx.
D . 2 π x . l = q, so that D = q/(2 π x)
F
i
g
u
r
e
3
.
3
4
-
A
n
a
l
o
g
o
u
s
h
e
a
t-
fl
o
w

a
n
d
c
h
a
r
g
e
-
fl
o
w

li
n
e
s
159 J.R.Lucas
High Voltage Cables
r
r
log
2
q
= x d .
x
1
.
2
q
= V
= where
2
1
e
r

r
r 0
1
2
ε π ε π
ε ε
ε
π ε
ξ


∴ ,
x 2
q
=
Considering the analogous heat flow case,
Letdθ = drop in temperature across dx
k = thermal resistivity
r
r
log
2
H k
= gives n integratio
x 2
dx . H . k
= d then
2
1
e
π
θ
π
θ
Thermal resistance is θ/H, so that
F
i
g
u
r
e
3
.
3
4
-
A
n
a
l
o
g
o
u
s
h
e
a
t-
fl
o
w

a
n
d
c
h
a
r
g
e
-
fl
o
w

li
n
e
s
160 J.R.Lucas
High Voltage Cables
r
r
log
2
k
= S resistance thermal
2
1
e
π

161 J.R.Lucas
High Voltage Cables
3.5.4 Thermal resistance of three-core cables
For three-core cables, the following two quip expressions are used.
dielectric of radius outer =
r

lies conductor each of centres he at which t radius = a
conductor of radius = r
where
r
a r
3
a
-
r
ln
6
k
= S (ii)
conductor of radius = r
insulation conductor of thickness = T
insulation belt of thickness = t
where
km ln
2
2
2
3
6
2
6
]
]
]


]
]
]

,
`

.
|

,
`

.
|

,
`

.
|
π
π
/ 1 +
r
t + T

T
t 1.1
+ 4.15
T
t 0.2
+ 0.85
6
k
= S (i)
162 J.R.Lucas
High Voltage Cables
3.5.5 Thermal resistance of protective coverings
Since the protective covering of the cable is in the form of a cylinder, the
expression is of the same form as that for a single core cable.
A/2 +
r
A/2 -
r
2
k
= S
2
3
ln
π
where
r
3
= radius of outer covering of cable
r
2
= radius of lead sheath
A = thickness of armouring
k = thermal resistivity
163 J.R.Lucas
High Voltage Cables
3.5.6 Thermal resistance of ground around cable
164 J.R.Lucas
H-H hearth Heat flow

−q
q h earth charge
flow
h
High Voltage Cables

F
i
g
u
r
e
3
.
3
5
-
E
f
f
e
c
t
o
f
E
a
r
t
h
S
u
r
f
a
c
e
165 J.R.Lucas
High Voltage Cables
Inside the cable constant temperature lines would all be concentric
cylinders since the outer lead sheath is a conductor of heat.
The flow of heat would consequently be radial.
However, outside the cable, the equi-temperature lines would no longer be
concentric and the heat flow would go radially outwards from the surface
of the cable and ending up at the surface of the ground normally
(assuming that the surface of the earth is at a constant temperature).
Here again, let us analyse using the analogy of the infinitely long
conductor carrying a charge q per unit length, placed at a distance h above
the earth surface (Figure 3.35).
This has the same effect as having a charge of -q at the same distance
beneath the earth.
The effect of the earth can be replaced by an equal and opposite charge on
the opposite side of the surface at the same distance from the surface.
F
i
g
u
r
e
3
.
3
5
-
E
f
f
e
c
t
o
f
E
a
r
t
h
S
u
r
f
a
c
e
166 J.R.Lucas
High Voltage Cables
The effects of the charge +q and -q can now be separately considered, and
the results superimposed.
Each charge considered separately will give rise to radial flux lines.
The potential difference between the charges caused by one of the charges
is given by
r
h 2

2
q
= V
e
log
ε π
The total potential difference caused between the charges is twice that of
the individual charge.
This is equal to
r
h 2

2
q 2
=
r
h 2

2
q) (- - q
= V
e e
log log
ε π ε π
Thus the potential difference to the neutral of each conductor is given as
167 J.R.Lucas
High Voltage Cables
r
h 2

2
q
= V
e
log
ε π
Analogy: The temperature difference from the heat source to earth is
given by
r
h 2

2
H k
=
e
log
π
θ
Thus the thermal resistance of the ground is given by
r
h 2
log
2
k
=
H
= resistance thermal
e
π
θ
168 J.R.Lucas
High Voltage Cables
When applied to the practical case, it is found that the theoretical thermal
resistance a found above has to be multiplied by a factor of 2/3.
This is because we have assumed the earth to be a plane of perfect
conductivity (or constant temperature).
Thus the modified thermal resistance G of the ground for practical
application is given by
r
h 2

3
k
= G
e
log
π
A representative value of the thermal resistivity k of the soil of average
moisture content is 180.
169 J.R.Lucas
High Voltage Cables
3.5.7 Cables exposed to air
The heat dissipation of a cable exposed to the air depends on the radiation.
For a surface in direct contact with the air, with unrestricted ventilation, the heat
dissipation is given by
r
with varies value hose constant w emissivity = k
, mperature ambient te =
, surface cable of e temperatur =
, sheath lead of usually cable, of radius exernal =
r

where
length of watt/cm
2
a
s
2
2
θ
θ
θ θ
π ) - ( kr 2 = H
1.25
a s
170 J.R.Lucas
High Voltage Cables
3.6High Voltage Bushings
Bushings are insulators which are used to take high voltage conductors
through earthed barriers such as walls, floors, metal, tanks etc.
The bushings have to provide electrical insulation of the conductor for the
working voltage and for various over-voltages which may occur in service
and also have to provide mechanical support against various mechanical
forces.
171 J.R.Lucas
High Voltage Cables
3.6.1 Simple cylindrical bushing
172 J.R.Lucas
bushi
ng
live
conductor
s
LtRrearthed
barrier
voltag
e
to
earth
V
r R
radius x
High Voltage Cables
F
i
g
u
r
e
3
.
3
6
-
S
i
m
p
l
e
c
y
li
n
d
ri
c
a
l
b
u
s
h
i
n
g
173 J.R.Lucas
High Voltage Cables
The simplest form of bushing is a cylinder of insulating material around
the conductor (Figure 3.36), with radial clearance t = (R - r) and axial
length L to suit electrical strengths of the insulating material and
surrounding media.
In this case, the voltage distribution with radius x is not linear so that the
material is not equally stressed.
The expressions for stress ξ and voltage to earth V
x
at radius x is given by
r
R
x
x
R
V
=
V
,
r
R
x
V
=
x
ln
ln
ln
ξ
The maximum stress in the bushing occurs at the conductor surface (i.e.
at x = r), so that
F
i
g
u
r
e
3
.
3
6
-
S
i
m
p
l
e
c
y
li
n
d
ri
c
a
l
b
u
s
h
i
n
g
174 J.R.Lucas
High Voltage Cables
r
t + r
r
V
=
e
log
max
ξ
so that the thickness t required for the bushing is given by
.... +
r 2
V
+
V
= t
1 -
r
V
r = t

2
2

ξ
ξ
ξ
max
max
max
exp
]
]
]
]

,
`

.
|
Thus as the voltage increases, the dimensions required become very large,
so much so that for very high voltages, simple cylindrical bushings of this
form are not satisfactory.
175 J.R.Lucas
High Voltage Cables
Upto about 66 kV, porcelain bushings (with or without oil) may be used.
176 J.R.Lucas
High Voltage Cables
3.6.2 Condenser bushing
The difficulty of the dimensions increasing rapidly is overcome by the
condenser bushing. In this case, the busing is divided into a number of
capacitors by concentric cylinders of metallic foil or metallic coated
paper.
By proper choice of lengths of these cylinders, it is possible to obtain a
nearly uniform voltage distribution with radius (Figure 3.37).
177 J.R.Lucas
earthed flange equipotential
cylinders
live conductor
High Voltage Cables













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
178 J.R.Lucas
High Voltage Cables
Since the stress is now made uniform, it must be equal to the ratio of the
applied voltage V to the thickness t' of the insulation.
ξ
ξ
max
max
V
= t that so


,
t
V
=
As can be seen when comparing with the simple bushing, the thickness t'
required now is much less than t.
Example
A condenser bushing for an r.m.s. voltage of 30 kV to earth is designed to
have a uniform radial voltage gradient (Figure 3.38).
The insulating material used has a maximum permissible working voltage
stress of 10 kV/cm (peak).
Assuming a uniform and very small thickness of insulation between each
successive foil, determine the radial thickness t' of the bushing.













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
179 J.R.Lucas
High Voltage Cables
If the length of the bushing at the outermost radius is 10 cm, determine the
length at the surface of the conductor (radius 2 cm).
Estimate also the thickness t for the bushing without foils, if it is to have
the same maximum radial stress.













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
180 J.R.Lucas
High Voltage Cables













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
l10 cm
Figure 3.38 = Length of condenser bushing
181 J.R.Lucas
High Voltage Cables
Since stress is uniform,
cm
max
4.24 =
10
2 30
=
V
= t
ξ

The profile of the bushing has the equation y = a/x,
at x = t' + 2, y = 10 cm, so that a = 10(t' + 2) = 10(4.24 + 2)
therefore, x.y = a = 62.4
at x = r = 2, y = l
therefore, l = 62.4/x = 62.4/2 = 31.2 cm
In the absence of foils,













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
182 J.R.Lucas
High Voltage Cables
cm 14.68 = 7.342 x 2 = t 8.342 = t + 1 . e . i
log
log log
2
1
max



2.121 = 2 1.5 = t) + (1

2
2 + t
2
2 30
= 10
r
t + r
r
V
=
2
1
e
e e
ξ
Thus in the absence of grading, it is seen that a much greater thickness of
insulation is required (14.68 cm as compared to 4.24 cm).
In addition to the simple cylinder bushing, and the condenser type
bushing, there are other types of bushings, which may consist of more
than one material.













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
4
Measurement of High Voltage













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
Measurement of High Voltages J.R. Lucas 184
4.0High Voltage Measurement
High voltages can be measured in a variety of ways. Direct measurement
of high voltages is possible up to about 200 kV, and several forms of
voltmeters have been devised which can be connected directly across the
test circuit.
High Voltages are also measured by stepping down the voltage by using
transformers and potential dividers.
The sparkover of sphere gaps and other gaps are also used, especially in
the calibration of meters in high voltage measurements.
Transient voltages may be recorded through potential dividers and
oscilloscopes.
Lightning surges may be recorded using the Klydonograph.













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
Measurement of High Voltages J.R. Lucas 185
4.1 Direct Measurement of High Voltages
4.1.1 Electrostatic Voltmeters
One of the direct methods of measuring high voltages is by means of
electro-static voltmeters.
For voltages above 10 kV, generally the attracted disc type of electrostatic
voltmeter is used.
When two parallel conducting plates (cross section area A and spacing x)
are charged q and have a potential difference V, then the energy stored in
the is given by













┌┐













││













││













││











┌─
╥─
─┘











──
──
─┴
─╨
──
─┴
──
──
──
──
──










──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──
──








──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──






──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──



──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──























──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──




──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──







──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──
──









──
──
──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──
──











──
──
──
──
──
──
──
──
──
──











──
──
┬─
╥─
──
┬─
──
──
──
──












└─
╨─
─┐














││













││













││













└┘
Fig
ure
3.3
7 -
Co
nde
nse
r
Bus
hin
g
Measurement of High Voltages J.R. Lucas 186
N
x
V
A - = F

x
A
- =
x d
C d
that so
x
A
= C e Capacitanc field uniform for

N
x d
C d

V
= F Force

x d F = C d
V
= W d change that so
V
C = W stored Energy
2
2
2
1
2
2
2
1
2
2
1 2
2
1
ε
ε ε


It is thus seen that the force of attraction is proportional to the square of
the potential difference applied, so that the meter reads the square value
(or can be marked to read the rms value).
Measurement of High Voltages J.R. Lucas 187
Electrostatic voltmeters of the attracted disc type may be connected across
the high voltage circuit directly to measure up to about 200 kV, without
the use of any potential divider or other reduction method.
[The force in these electrostatic instruments can be used to measure both
a.c. and d.c. voltages].
k
V
sprin
g
l.v.
electrod
e
h.v.
electrode
markings on x-
axis
Figure 4.1 – Abraham
electrostatic voltmeter
Measurement of High Voltages J.R. Lucas 188
Abraham Voltmeter
The Abraham voltmeter is the most commonly used electrostatic meter in
high voltage testing equipment.
There are two mushroom shaped hollow metal discs.
The right hand electrode forms the high voltage plate, while the centre
portion of the left hand disc is cut away and encloses a small disc which is
movable and is geared to the pointer of the instrument.
The range of the instrument can be altered by setting the right hand disc at
pre-marked distances.
The two large discs form adequate protection for the working parts of the
instrument against external electrostatic disturbances.
These instruments are made to cover ranges from 3 kV to 500 kV.
Owing to the difficulty of designing electrostatic voltmeters for the
measurement of extra high voltages which will be free from errors due to
corona effects, within the instrument, and to the external electrostatic
fields, a number of special methods have been devised for the purpose.
Measurement of High Voltages J.R. Lucas 189
4.1.2 Sphere gaps
The sphere gap method of measuring high voltage is the most reliable and
is used as the standard for calibration purposes.
The breakdown strength of a gas depends on the ionisation of the gas
molecules, and on the density of the gas.
As such, the breakdown voltage varies with the gap spacing; and for a
uniform field gap, a high consistency could be obtained, so that the
sphere gap is very useful as a measuring device.
By precise experiments, the breakdown voltage variation with gap
spacing, for different diameters and distances, have been calculated and
represented in charts.
In measuring device, two metal spheres are used, separated by a gas-gap.
The potential difference between the spheres is raised until a spark passes
between them.
Measurement of High Voltages J.R. Lucas 190
The breakdown strength of a gas depends on the size of the spheres, their
distance apart and a number of other factors.
A spark gap may be used for the determination of the peak value of a
voltage wave, and for the checking and calibrating of voltmeters and other
voltage measuring devices.
The density of the gas (generally air) affects the spark-over voltage for a
given gap setting. Thus the correction for any air density change must be
made. The air density correction factor δ must be used.
]
]
]

×
t + 273
P
0.386 =
t + 273
20 + 273
760
P
= δ
The spark over voltage for a given gap setting under the standard
conditions (760 torr pressure and at 20
o
C) must be multiplied by the
correction factor to obtain the actual spark-over voltage.
Measurement of High Voltages J.R. Lucas 191
The breakdown voltage of the sphere gap (figure 4.2) is almost
independent of humidity of the atmosphere, but the presence of dew on
the surface lowers the breakdown voltage and hence invalidates the
calibrations.
Measurement of High Voltages J.R. Lucas 192

where d = gap spacing, D = sphere diameter
dD
F
i
g
u
r
e
4
.
-
M
e
a
s
u
ri
n
g
s
p
h
e
r
e
s
Figure 4. - Breakdown voltage
characteristic of sphere gaps
Measurement of High Voltages J.R. Lucas 193
The breakdown voltage characteristic (figure 4.3) has been determined for
similar pairs of spheres (diameters 62.5 mm, 125 mm, 250 mm, 500 mm,
1 m and 2 m)

When the gap distance is increased, the uniform field between the spheres
becomes distorted, and accuracy falls.
The limits of accuracy are dependant on the ratio of the spacing d to the
sphere diameter D, as follows.
d< 0.5 D, accuracy = ± 3 %
0.75 D> d > 0.5 D, accuracy = ± 5 %
F
i
g
u
r
e
4
.
-
M
e
a
s
u
ri
n
g
s
p
h
e
r
e
s
gap
distance
breakdow
n voltage
Measurement of High Voltages J.R. Lucas 194
For accurate measurement purposes, gap distances in excess of 0.75D are
not used.
The breakdown voltage characteristic is also dependant on the polarity of
the high voltage sphere in the case of asymmetrical gaps (i.e. gaps where
gap spacing
(mm)
breakdown
voltage
(kV)
Figure 4.4 – Breakdown voltage
characteristics
d < D

+h.v.

-h.v. &
a.c.
DD2D D/
10
Measurement of High Voltages J.R. Lucas 195
one electrode is at high voltage and the other at a low voltage or earth
potential).
If both electrodes are at equal high voltage of opposite polarity (i.e. + ½ V
and - ½ V), as in a symmetrical gap, then the polarity has no effect.
Figure 4.5 -
sphere gap
Measurement of High Voltages J.R. Lucas 196
In the case of the asymmetrical gap, there are two breakdown
characteristics; one for the positive high voltage and the other for the
negative high voltage.
Since the breakdown is caused by the flow of electrons, when the high
voltage electrode is positive, a higher voltage is generally necessary for
breakdown than when the high voltage electrode is negative.
However, when the gaps are very far apart, then the positive and the
negative characteristics cross over due to various space charge effects.
But this occurs well beyond the useful operating region.
Under alternating voltage conditions, breakdown will occur corresponding
to the lower curve (i.e. in the negative half cycle under normal gap
spacings).
Thus under normal conditions, the a.c. characteristic is the same as the
negative characteristic.
Measurement of High Voltages J.R. Lucas 197
In sphere gaps used in measurement, to obtain high accuracy, the
minimum clearance to be maintained between the spheres and the
neighbouring bodies and the diameter of shafts are also specified, since
these also affect the accuracy (figure 4.5).
There is also a tolerance specified for radius of curvature of the spheres.
"The length of any diameter shall not differ from the correct value by
more than 1% for spheres of diameter up to 100 cm or more than 2% for
larger spheres".
Peak values of voltages may be measured from 2 kV up to about 2500 kV
by means of spheres.
One sphere may be earthed with the other being the high voltage
electrode, or both may be supplied with equal positive and negative
voltages with respect to earth (symmetrical gap).
When spark gaps are to be calibrated using a standard sphere gap, the two
gaps should not be connected in parallel.
Measurement of High Voltages J.R. Lucas 198
Equivalent spacing should be determined by comparing each gap in turn
with a suitable indicating instrument.
Needle gaps may also be used in the measurement of voltages up to about
50 kV, but errors are caused by the variation of the sharpness of the
needle gaps, and by the corona forming at the points before the gap
actually sparks over.
Also the effect of the variation of the humidity of the atmosphere on such
gaps is much greater.
Usually, a resistance is used in series with the sphere gap, of about 1 Ω/V
so as to limit the current under sparkover conditions to about a maximum
of 1 A.
However for impulse measurements, a series resistance must not be used
since this causes a large drop across the resistance.
Measurement of High Voltages J.R. Lucas 199

In measuring impulse voltages, since the breakdown does not occur at
exactly the same value of voltage each time, what is generally specified is
the 50 % breakdown value.
A number of same value impulses is applied and a record is kept of the
number of times breakdown occurs, and a histogram is plotted with the
peak value of impulse voltage and percentage of breakdown (figure 4.6).
Percent
age
breakd
own
Figure 4.6 - Breakdown voltage
characteristic for impulses
100
%
50
%
breakdown
voltage (peak)
Measurement of High Voltages J.R. Lucas 200
4.2Transformer and potential divider methods of measurement
4.2.1 Transformer ratio method
Tes
t
devi
ce
Curr
ent
limit
ing
resi
stor
1
Ω /V
a.c.
sup
ply
V
Figure 4. -
transformer ratio
method
Measurement of High Voltages J.R. Lucas 201
The use of primary voltage to estimate secondary voltage is a fairly rough
method of measurement, but is satisfactory enough for most ac tests.
In this method (figure 4.7), the voltage on the low voltage side of the
high-tension transformer is measured.
The actual voltage across the load is not measured.
Since the current taken by the device under test is usually very small,
currents such as due to corona may cause considerable error in the
measured voltage.
This method measures the rms voltage. In order to determine the peak
value it is necessary to determine the wave form of the secondary voltage.
Test
devi
ce
Curr
ent
limiti
ng
resis
tor
1
Ω/V
a.c.
sup
ply
V
Figure 4.8 - with additional
potential winding
Measurement of High Voltages J.R. Lucas 202
Some high voltage transformers (figure 4.8) carry a separate voltmeter-
coil having a number of turns which is a definite fraction of the secondary
turns. This method cannot be used with the cascade arrangement of the
transformers.
It may also be possible to have a potential transformer connected across
the test device and the voltage measured, however this is an expensive
arrangement (figure 4.9).
Measurement of High Voltages J.R. Lucas 203

Test
devic
e
Curre
nt
limiti
ng
resist
or
1
Ω/V
a.c.
suppl
y
V
Figure 4.9 - with potential
transformer
Range
0 –
100 V
potential
transformer
h.v.
transformer
Measurement of High Voltages J.R. Lucas 204
Even this method may not be very satisfactory under very high voltage
conditions and the series resistance method of measurement may be used.
Measurement of High Voltages J.R. Lucas 205
Series resistance method of measurement
In the series resistance method a high series resistance (specially designed
to withstand high voltage) and resistance of 20 kΩ/V, is used with micro-
ammeter (having a 50 μA movement).
Method is applicable for both ac and dc.
A number of resistances would be necessary in series, and to prevent
leakage current, we would have to have the whole system in a insulated
container, which is earthed for shielding purposes.
Figure 4. - Series
resistor microammeter
μ A
Safety
gap
high
series
resista
nce
(20
kΩ/V)
50 μA

move
ment
h.
v.
Measurement of High Voltages J.R. Lucas 206
As a safety measure, a safety gap or neon lamp is connected across the
micro-ammeter.
With a stable supply (accuracy 0.10%) final accuracy of 1% is achieved.
Measurement of High Voltages J.R. Lucas 207

Test
devi
ce
Curr
ent
limiti
ng
resis
tor
1
Ω /V
a.c.
sup
ply
V
Figure 4.11 - Resistive
potential divider method
resistive
divider
(≈ 20
kΩ/V)
h.v.
transforme
r
Measurement of High Voltages J.R. Lucas 208
When the above method is used for alternating voltages, there would be
the effect of the distributed capacitances as well.
The capacitive effects can be reduced by providing a suitable screen, or by
balancing the capacitance.
Measurement of High Voltages J.R. Lucas 209
4.2.2 Resistive potential divider method
In this method, a high resistance potential divider is connected across the
high-voltage winding, and a definite fraction of the total voltage is
measured by means of a low voltage voltmeter.
Under alternating conditions there would be distributed capacitances.
Measurement of High Voltages J.R. Lucas 210

One method of eliminating this would be to have a distributed screen
of many sections and using an auxiliary potential divider to give
fixed potential to the screens.
Figure 4. - Screening of resistive dividers (a) (b)
Measurement of High Voltages J.R. Lucas 211
The currents flowing in the capacitances would be opposite in directions
at each half of the screen so that there would be no net capacitive current
(Figure 4.12 (a)).
It also possible to have a metal conical screen (Figure 4.12 (b)).
The design has to be done by trial and error.
There would be capacitances to the conical screen as well as capacitances
to earth, so that if at any point the capacitive current from conical screen
to the point is equal to that from the point to the earth, then the
capacitances would have no net effect.
Measurement of High Voltages J.R. Lucas 212
4.2.3 Capacitive potential divider method
Measurement of High Voltages J.R. Lucas 213
Test
devic
e
Curre
nt
limiti
ng
resist
or
1
Ω /V
a.c.
suppl
y
V
Figure 4. - Capacitive potential
divider method
h.v.
capacitor
h.v.
transformer
stray
capaci
tors
electrost
atic
voltmete
r
C
1
C
2
Measurement of High Voltages J.R. Lucas 214
For alternating work, instead of using a resistive potential divider, we
could use a capacitive potential divider.
In this two capacitances C
1
and C
2
are used in series, the electrostatic
voltmeter being connected across the lower capacitor.
If the system is kept at a fixed position, we can make corrections for the
fixed stray capacitances.
Or if screens are used, the capacitance to the screen would be a
constant,and we could lump them up with the capacitances of the arms.
Neglecting the capacitance of the voltmeter (or lumping the electrostatic
voltmeter capacitance with C
2
) the effective capacitance of C
1
and C
2
in
series is C
1
C
2
/(C
1
+C
2
), and since the charge is the same,
V .
C
+
C
C
= V .
C C
/ )
C
+
C
(
C
/ 1
=
C
across Voltage
2 1
1
2 1 2 1
2
2
Measurement of High Voltages J.R. Lucas 215
The capacitance of h.v. standard capacitor must be accurately known, and
the capacitance must be free from dielectric losses.
For this reason, air capacitances are always used for this purpose.
This method also measures the r.m.s. value.
It is sometimes more useful to have a measure of the peak value of the
alternating voltage rather than the r.m.s. value,since it is the peak value of
the applied voltage which produces the actual breakdown stress in the
material under test.
If the shape of the voltage waveform is known, the peak voltage may be
obtained from the r.m.s. voltage.
It is often more satisfactory however, to use some method of voltage
measurement which gives the peak value of the voltage directly.
Measurement of High Voltages J.R. Lucas 216
4.2.4 Matching of Potential dividers
When waveforms are observed on the oscilloscope, through a potential
divider, a cable is necessary to connect the test waveform to the
oscilloscope, and also to cause a small delay between the arrival of the
trigger pulse and the waveform (Figure 4.14).
Measurement of High Voltages J.R. Lucas 217

cable
C
2
E
1
E
2
E
3
E
4
Z
0
Y-
plate
s
input
impul
se
Device
under
test
Tim
e
base
trigger pulse single sweep X-
plate
s
Figure 4. - Observation of impulse waveform through
potential divider
Measurement of High Voltages J.R. Lucas 218
If the delay cable is lossless , then it may be represented by purely
inductances and capacitors, so that the surge impedance of cable or delay
network = Zo = [z/y]
1/2
.
In a lossless cable , Zo is purely resistive.
Velocity of the wave in cable =
ε
ε µ
r
8
0
10
3
=

1

C L
1 ×

The oscilloscope can display a maximum of about 50 V to 100 V and thus
the impulse voltage must be reduced by a suitable potential divider.
The requirement of the potential divider used are that it reduces the
applied voltage without producing any distortion (i.e: the ratio of the
potential divider does not vary with time or frequency).
The potential divider can be of two types.
(i) Resistive
Measurement of High Voltages J.R. Lucas 219
and (ii)Capacitive.
In practice , neither case is obtained in the pure form, but a mixture of
both.
The capacitive effect in the resistive divider is much more than the
resistive effect on the capacitive divider.
In the case of the resistive divider , the lower arm of the divider has its
resistance fixed by the surge impedance of the cable used (for matching )
and by the wave–tail requirements of the impulse generator circuit (if any
impulse generator is used).
The ratio of the divider is determined by the sensitivity of the C.R.O and
the voltage.
The capacitive divider is generally bulkier than the resistive divider , but
has several advantages.
It can be used as part of the wavefront forming circuit.
Measurement of High Voltages J.R. Lucas 220
The self capacitance of the cable connecting the device to the C.R.O adds
to the capacitance of the cable connecting the divider to the oscilloscope
adds to the capacitance of the lower arm.
When the initial part of the surge enters the cable, it acts as a transmission
line and presents its surge impedance to the surge, but when the line
becomes charged, it behaves as a capacitor.
When using potential dividers, it is necessary to suitably terminate the
cable at the two ends so as to have perfect matching of the cables at two
ends.
Measurement of High Voltages J.R. Lucas 221
Resistive potential dividers
There are three ways in which they may be matched to delay cables.
(1) Matching at potential divider end only:
Measurement of High Voltages J.R. Lucas 222
R
1
R
2
cable
C
2
E
1
E
2
E
3
E
4
R
3
Z
0
Y-
plate
s
input
impu
lse
Device
under
test
Ti
me
bas
e
trigger
pulse
single
sweep
X-
plate
s
Figure 4.15 – Matching of resistive
divider at sending end only
R
1
R
2
R
3
Measurement of High Voltages J.R. Lucas 223
In this arrangement, the receiving end (ie: end connected across CRO
Y-plates) is kept on open circuit.
We try to obtain perfect matching at the sending end so that there is no
reflection, and perfect reflection at the receiving end.
For perfect matching at the sending end,
the equivalent impedance of the section before the cable must be Z
O

Impedance = R
3
+ R
1
//R
2
= R
3
+ R
1
.R
2
/(R
1
+ R
2
)
= Z
O
for perfect matching at s.e
Figure
4.16
Measurement of High Voltages J.R. Lucas 224
If R
1
>> R
2
as usually is,
then we have R1//R2 ≈ R2 ,
Z
O
= R
2
+ R
3
At the junction of the divider E
2
, the equivalent impedance to earth Z
1
is
given by
Measurement of High Voltages J.R. Lucas 225
E
.
)
R
+
Z
( 2
R
=
E
.
)
R
+
Z
(
Z
2
)
Z
+
R
(
R
.
)
Z
+
R
(
Z
=
E
.
Z
+
R
Z
=
E
that so

E
.
)
R
+
Z
(
Z
2
)
Z
+
R
(
R
=
E
.
R
+
Z
Z
=
E
junction at voltage
1
1 1
2
1
1 1 0
0 3 2
0 3
0
2
0 3
0
3
1
1 1 0
0 3 2
1
1 1
1
2


R
+
R
=
Z
,
Z
2
)
Z
+
R
(
R
=
)
Z
+
R
+
R
(
)
Z
+
R
(
R
= )
Z
+
R
( / /
R
=
Z 3 2 0
0
0 3 2
0 3 2
0 3 2
0 3 2 1

This voltage waveform E
3
travels towards the receiving end and is
reflected at the open end without change of sign, so that the voltage
transmitted to the CRO is 2 E
3
.
Measurement of High Voltages J.R. Lucas 226
Z
2
)
Z
+
R
(
R
=
Z
where
0
0 3 2
1
,
E
.
R
+
Z
R
=
E
=
E

1
1 1
2
4 r

OR
If the lower arm itself is balanced,
i.e. R
2
=Z
0
,
then R
3
= 0
and the voltage transmitted to the oscilloscope is given by
E
.
Z
+
R
Z
=
E 1
0
2
1
1
0
r
Measurement of High Voltages J.R. Lucas 227
(2) Matching the cable at the oscilloscope end only:
Measurement of High Voltages J.R. Lucas 228
R
1
R
2
cable
C
k
E
1
E
2
E
3
E
4
R
4
Z
0
Y-
plate
s
input
impu
lse
Device
under
test
Ti
me
bas
e
trigger
pulse
single
sweep
X-
plate
s
Figure 4. - Matching of resistive
divider at receiving end only
Measurement of High Voltages J.R. Lucas 229
In this arrangement, the cable is matched only at the receiving end so that
there will no reflection at this end.

E
.
R
+
Z
Z
=
E

Z
+
R
Z R
=
Z
/ /
R
=
Z
= 2 E at impedance Equivalent
1
1 1
1
2
0 2
0 2
0 2 1

Since the cable is properly matched at the receiving end, R
4
= Z
0
The voltage wave E
2
travels along the cable, and since there is proper
matching at the receiving end, it is transmitted with out any reflection.
E
.
R R
+
Z
)
R
+
R
(
Z R
=
E
.
R
)
Z
+
R
( + +
Z R
Z R
=
E
.
R
+
Z
Z
=
E
=
E

1
2 1 0 2 1
0 2
1
1 0 2 0 2
0 2
1
1 1
1
2 4

For given values of R
1
, R
2
and E
1
this arrangement gives smaller voltages
at the C.R.O than when only the divider end is matched.
Measurement of High Voltages J.R. Lucas 230
If the point E
2
is not connected to earth through the resistance R
2
,
(ie. if R
2
= ), then we have
Measurement of High Voltages J.R. Lucas 231
R
1
cable
C
k
E
1
E
2
E
3
E
4
R
4
Z
0
Y-
plate
s
input
impul
se
Device
under
test
Tim
e
bas
e
trigger
pulse
single
sweep
X-
plate
s
Figure 4. - Matching of resistive divider at receiving
end only (with R
2
= ∞)
Measurement of High Voltages J.R. Lucas 232
For this case the voltage at the oscilloscope is given by
Measurement of High Voltages J.R. Lucas 233
E
.
R
+
Z
Z
=
E
.
R R
+
Z
)
R
+
R
(
Z R
=
E 1
1 0
0
1
2 1 0 2 1
0 2

R
4
Lt
2
∞ →
(3) Matching at both ends
of the cable :
Measurement of High Voltages J.R. Lucas 234

R
1
R
2
cable
C
k
E
1
E
2
E
3
E
4
R
4
R
3
Z
0
Y-
plate
s
input
impu
lse
Device
under
test
Ti
me
bas
e
trigger
pulse
single
sweep
X-
plate
s
Figure 4. - Matching of resistive
divider at both ends
Measurement of High Voltages J.R. Lucas 235
In this case, the cable is matched at both ends .
With this termination there is no reflection at either end.
This arrangement is used when it is necessary to reduce to a minimum the
irregularities produced in the delay cable circuit.
As before, for perfect matching at receiving end, R
4
= Z
0
and for perfect matching at sending end, R
2
+ R
3
= Z
0
.
Also, at E
2
, the equivalent impedance Z
1
to earth is given by
Measurement of High Voltages J.R. Lucas 236
E
.
)
R
+
Z
( 2
R
=
E
.
)
R
+
Z
(
Z
2
)
Z
+
R
(
R
.
)
Z
+
R
(
Z
=
E
.
Z
+
R
Z
=
E
that so

E
.
)
R
+
Z
(
Z
2
)
Z
+
R
(
R
=
E
.
R
+
Z
Z
=
E
junction at voltage
1
1 1
2
1
1 1 0
0 3 2
0 3
0
2
0 3
0
3
1
1 1 0
0 3 2
1
1 1
1
2


R
+
R
=
Z
,
Z
2
)
Z
+
R
(
R
=
)
Z
+
R
+
R
(
)
Z
+
R
(
R
= )
Z
+
R
( / /
R
=
Z 3 2 0
0
0 3 2
0 3 2
0 3 2
0 3 2 1

Due to perfect matching at the receiving end, this is transmitted without
any reflections.
E
.
)
Z
+
R
( 2
R
=
E
=
E

1
1 1
2
3 4

t
Measurement of High Voltages J.R. Lucas 237

Figure 4. - Stray capacitances and chopped wave s
tr
a
y
c
a
p
a
c
it
a
n
c
e
s





c
h
o
p
p
e
d
w
a
v
e
Measurement of High Voltages J.R. Lucas 238
The stray capacitances present between the turns of the resistances would
make the current distribution along the resistance non–uniform.
When the rate of change of voltage is high, then the errors due to the
capacitances are large (especially in waves such as chopped waves).
By having a distributed capacitance along the resistance which are larger
than the stray capacitances, the effect of the stray capacitance may be
eliminated.
An easier way to compensate for the stray capacitances is by having
capacitive potential divider instead of the resistive divider.
Capacitive potential dividers
The effect of stray capacitance may be made constant, in a capacitive
divider, by shielding the potential, divider; and hence make an allowance
for it.
s
tr
a
y
c
a
p
a
c
it
a
n
c
e
s





Measurement of High Voltages J.R. Lucas 239
The disadvantage of the capacitive potential divider is that proper
termination cannot be done.
There are two methods used to couple capacitive dividers to delay cables.
Measurement of High Voltages J.R. Lucas 240
(1) Simple capacitor connection
Measurement of High Voltages J.R. Lucas 241

C
1
C
2
C
k
cab
leC
2
E
1
E
2
E
3
E
4
R
1
=
Z
0
Z
0
Y-
plate
s
input
impu
lse
Figure 4. - Matching of capacitive divider
(simple capacitor connection)
trigg
er
Measurement of High Voltages J.R. Lucas 242
In the simple capacitor connection, we attempt to prevent reflections at
sending end.
The sending end is terminated with a resistance R
1
= Z
o
in series with the
cable.
Initially the cable capacitance would not have charged up, and only C
1
and
C
2
would be present.
Initially,
Z
=
R
matching for ,
E
=
R
+
Z
Z
.
E
=
E
C
+
C
C
.
E
=
E
0 1 2
2
1
1 0
0
2 3
2 1
1
1 2
Due to perfect reflection at the receiving end, E
3
travelling towards it
would be reflected and hence the voltage transmitted to the CRO would be
doubled.
Figure 4. - Waveforms for simple
capacitor connection
Measurement of High Voltages J.R. Lucas 243
C
+
C
C
.
E
=
E
=
E
x 2 =
E
2 =
E
. e . i
2 1
1
1 2 2
2
1
3 r
This gives the amplitude of the voltage wave as it reaches the Y–plates.
As time goes on, the cable capacitance charges up and behaves as a
capacitance in parallel with the lower arm.
Therefore, after infinite time, voltage at Y–plates would be given by
E
4
ti
m
e

delay

Measurement of High Voltages J.R. Lucas 244
C
+
C
+
C
C
.
E
=
E
k 2 1
1
1 r
Thus the ratio of the input voltage to the output voltage of the capacitive
divider varies with time and we get a distorted output waveform displayed
on the oscilloscope.
Thus the capacitive potential divider introduces distortion.
The difference between the initial and final ratios will be appreciable
unless C
2
is at least 10 times that of the cable capacitance C
k
, in which
case the error would be about 10%.
This error can be reduced by transferring part of the low voltage capacitor
to the C.R.O. end of the delay cable and connecting it in series with a
resistance equal to the cable surge impedance Z
0
(resistive if cable is
lossless).
(2) Split capacitor connection
Measurement of High Voltages J.R. Lucas 245
C
1
C
2
R
1
E
1
E
2
E
3
E
4
Z
o
cable
C
k
C
3
R
2
y -
plates
x -
plates
Input
impulse
Figure 4. - Matching of capacitive divider (simple capacitor
connection)
Measurement of High Voltages J.R. Lucas 246
In this connection, in addition to matching the cable at the sending end (R
1
= Z
0
), it is also matched at the oscilloscope end (R
2
= Z
0
).
Further to ensure that the long term ratio remains the same as the initial
ratio, the lower end capacitor is split into C
2
and C
3
.
Initially the capacitances C
k
and C
3
would not have charged, and only the
capacitances C
1
and C
2
would be effective in the voltage ratio.
C
+
C
C
.
E
=
E
Initially
2 1
1
1 2
matching for
Z
=
R
,
E
=
Z
+
R
Z
.
E
=
E
also
0 1 2
2
1
0 1
0
2 3
Due to perfect matching at the receiving end, the voltage wave is
transmitted without any reflection.
Measurement of High Voltages J.R. Lucas 247
Therefore the observed voltage is given by
C
+
C
C
.
2
E
=
E
=
E
=
E

2 1
1 1
2
2
1
3 r

After infinite time, the capacitances C
k
and C
3
would have completely
charged up, and the receiving end in effect would be on open circuit, since
C
3
would no longer be conducting.
Measurement of High Voltages J.R. Lucas 248
Since all the capacitors C
2
, C
3
and C
k
are in parallel,
C
1
C
2
C
3
C
k
cab
leC
2
E
1
E
2
E
3
E
4
R
2
R
1
Z
0
Y-
plate
s
input
impu
lse
Figure 4. - Waveforms for split capacitor
connection
Measurement of High Voltages J.R. Lucas 249
C
+
C
+
C
+
C
C
.
E
=
E
k 3 2 1
1
1 2
If the initial and the final values of the ratio are made equal, then the
distortion is reduced to a great degree.
C
+
C
=
C
+
C
. e . i
k 3 2 1
C
+
C
+
C
+
C
C
.
E
=
C
+
C
C
.
2
E
k 3 2 1
1
1
2 1
1 1
If this condition is satisfied, then the distortion is low and near
faithful reproduction can be expected as shown in figure 4.25.
E
4
tim
e
→delay

Measurement of High Voltages J.R. Lucas 250
4.3Measurement of Surges
4.3.1 Klydonograph
Lightning is probably the most spectacular of the high voltage
phenomena.
Measurement of High Voltages J.R. Lucas 251
Very little is known about lightning, as it is not possible to create
lightning or to obtain a lightning strike when and where we please.
Also very little is known of its effects and the voltages of the surges that
appear in the transmission lines due to it.
The phenomena of the lightning could be studied to a certain extent by the
surges it produces on the transmission lines.
The frequency of occurrence of surge voltages and the magnitude of the
surge it produces on the transmission lines could be studied using
Litchenberg patterns obtained by using a Klydonograph.
metal
electrode
h.v.
electrod
e
li
ne
Figure 4. -
Klydonograph
dielec
tric
photograp
hic film
Measurement of High Voltages J.R. Lucas 252
The Klydonograph (Figure 4.26) has a dielectric sheet, on the surface of
which is placed a photographic film.
The insulator material separates a plane electrode on one side, and a
pointed electrode which is just in contact with the photographic film.
The high voltage is applied to the pointed electrode and the other
electrode is generally earthed.
The photographic film can be made to rotate continuously by a clockwork
mechanism.
The apparatus is enclosed in a blackened box so as not to expose the
photographic film.
When an impulse voltage is applied to the high voltage electrode, the
resultant photograph shows the growth of filamentary streamers which
develop outwards from the electrode.
Measurement of High Voltages J.R. Lucas 253
This imprint on the photographic plate is not due to normal photographic
action, and occurs even through there is no visible discharge between the
electrodes.
If flashover of the insulator or a visible discharge occurs, then the film
would become exposed and no patterns would be obtained.
These patterns obtained on the photographic film are known as
Litchenberg patterns.
When a positive high voltage is applied to the upper electrode, clearly
defined steamers which lie almost within a definite circle is obtained.
volt
age
(kV)
for
negativ
e
patterns
for
positive
patterns
radius of
pattern (mm)
h. v.
conductor
to
Klydonogr
aph
Measurement of High Voltages J.R. Lucas 254
If the voltage applied is negative, then the observed pattern is blurred and
the radius of the pattern is much smaller.
For both types of surges, the radius of the pattern obtained increases with
increase in voltage.
For a given apparatus with a fixed thickness of dielectric, the radius of the
pattern obtained (Figure 4.27a) is a definite function of the voltage
applied.
Thus by calibrating the Klydonograph using a high voltage oscilloscope
and known surge voltages, it is possible to use this apparatus to record
surges that occur.
Figure 4. -
Klydonograph
Measurement of High Voltages J.R. Lucas 255
If the positive voltage applied is increased beyond a certain value,
branching may occur along the branches coming out from the electrode.
The maximum voltage that can be measured using a Klydonograph is
dependant on the thickness of the dielectric material.
Thus to measure voltages beyond this value,such as occuring in
transmission lines, an insulator string potential divider is used. (Figure
4.27b)
For a fixed apparatus, for a positive high voltage applied as the top
electrode, the variation of the applied voltage with radius of the pattern
obtained is quite definite and the radius is quite large.
In the case of the negative high voltages, the characteristics is much more
variable and the radius is much smaller.
Thus usually it is preferable to use the positive pattern for the
measurement of high voltage surges.
Measurement of High Voltages J.R. Lucas 256
The applied voltage versus radius of pattern characteristics of the
Litchenberg pattern is shown in figure 4.28.
Since the surges due to lightning may be either positive or negative, and
since it is preferable to observe the positive pattern in either case, we
make a modification to the apparatus.
negati
ve
voltag
e
small
positi
ve
voltag
e
positi
ve
voltag
e
large positive
voltage
Figure 4. -
Litchenberg patterns
Measurement of High Voltages J.R. Lucas 257


In the modification shown in figure 4.29, there are two such instruments,
with the electrode connections made in opposite directions.
Thus in the modification, if a positive surge comes, then a positive pattern
would be recorded in (a), and a negative pattern in (b), of which the
pattern on (b) can be used for the measurement of the positive surge.
Negative Pattern Positive Pattern
Measurement of High Voltages J.R. Lucas 258
In the case of a negative surge, the opposite would happen, and the pattern
earth
electrode
h.v.
electrode
earth
electrode
h.v. electrode lin
e
Figure 4. - Klydonograph for measurement
of both polarities
(b
)
(a)
Measurement of High Voltages J.R. Lucas 259
Thus the magnitude of the surge as well as the polarity could be
determined from the Litchenberg patterns on (a) and (b).
Since the photographic film is continuously moving, it is possible in some
elaborate apparatus to record the date and time occurrence of the surge as
well.
Measurement of High Voltages J.R. Lucas 260
4.4General measurements
4.4.1 Peak reading voltmeters
(i) Capacitor charging method
tim
e
high
resista
nce
R
Test
devi
ce
Curr
ent
limit
ing
resis
tor
1
Ω/V
a.c.
sup
ply
C dio
de
Figure 4. - Capacitor
charging method
Measurement of High Voltages J.R. Lucas 261
In the positive half cycle, capacitor charges up to peak value, and when the
voltage falls it discharges (very slightly) through the milliammeter, so that the
voltage across the capacitor is very nearly a constant at the peak value and the
current is thus proportional to the peak value.
(Time constant RC must be very high in comparison to period of applied voltage).
(ii) Using neon lamp
A neon tube (if the voltage at which the lamp strikes is known) can be
used with a capacitive potential divider to obtain the peak value of an
applied voltage waveform.
suppl
y
l.v.
capacit
or
(variabl
e)
h.v.
capacito
r (fixed)
Figure 4. - Measurement
using neon lamp
neo
n
lam
p
Measurement of High Voltages J.R. Lucas 262
The low voltage variable capacitor is varied until the neon lamp strikes.
From the ratio of the capacitances, the supply voltage can be calculated.
Since the extinction voltage is more constant than the striking voltage, the
extinction voltage could be used as the standard.
An accuracy of ± ½ % could be obtained with the striking voltage and an
accuracy of ± ¼ % could be obtained with extinction voltage.
(iii) Rectifier-Capacitor current method
d.c.
meter
supply
V
h.v.
capacito
r
dio
de
dio
de
i+–
b
Figure 4. - Rectifier-capacitor
current method
Measurement of High Voltages J.R. Lucas 263
The best known and the most usual method of measuring the peak value is
the rectified capacitor current method.
A high voltage capacitor is connected to the hv supply with a rectifier
ammeter in the earth connection.
The indicated value will correspond to the peak value of the positive or
negative half cycle.
The diode used in series with the milliammeter should have a low forward
resistance and a high reverse resistance a ratio of 1:10
5
is desirable.
Silicon diodes provide an ideal rectifier for the purpose.
Measurement of High Voltages J.R. Lucas 264

BABAiviv
τ
V
m
ax
tttt(a) Sinusoidal
Waveform
(b) non-sinusoidal
Waveform
Figure 4. - Waveforms for
peak measurement
Supp
ly
d.c. +

Measurement of High Voltages J.R. Lucas 265
f C 2
I
=
C 2
I
.
=
V
giving ,
V
2 . C = .
I

V d . C = t d . i that so
av av
max max av
B
A
B
A
τ
τ ∴
∫ ∫
,
t d
v d
C = i
Figure 4. - Full wave
circuit
Measurement of High Voltages J.R. Lucas 266
Since a d.c meter is used it would read I
av
, and hence would correspond to
the maximum value of voltage, independent of the waveform, except in
the case when there is more than one maxima and minima per cycle.
In such a case the meter reading would no longer corresponds to the actual
maxima (Figure 4.33(b)), but an addition of successive peak-to-peaks.
Instead of using a half wave rectifying unit as in figure 4.32, we could
also used a full wave rectifying unit as shown in figure 4.34.
In this case, the reading of the meter would effectively be double giving
the result
f C 4
I
=
V
av
max
Measurement of High Voltages J.R. Lucas 267
Thus using either half wave or full wave rectifying units, we can obtain
the peak value of the voltage independent of the wave form, if the
capacitance and frequency are known from the reading of the d.c meter.
Measurement of High Voltages J.R. Lucas 268
4.4.2 Oscilloscope for measurement of fast transients
High voltage oscilloscopes are used for the study of fast transient
phenomena, particularly in the work on high-voltage and on spark
breakdown in small gaps.
These have a high sweep speed.
Since the speed is high, the intensity is lowered and hence a higher
intensity is required.
In these the beam should not come on till the transient comes in because;
(a) if it is stationary, the spot of high intensity would fog the photograph
before the transient comes on, and
(b) if it is moving, the beam may have swept before the transient comes.
Thus the beam should be brought on just before the transient comes on, by
being triggered by the transient.
Measurement of High Voltages J.R. Lucas 269
The transient should come on the Y-plate only shortly after the beam, so
that the whole transient is clearly seen.
For this a delay cable is used.
The delay cable ensures that the transient appears slightly after the beam
comes on.
Such a scope can have a maximum of 50 V to 100 V applied across Y
plate so that we would have to use a potential divider.
For high writing speed,the anode – cathode voltage should be high (50 –
100 kV).
The sweep generator should produce a single sweep (not repetitive), as
transients are not repetitive, and triggered by the signal.
The delay cable causes the signal to appear at the Y–plates a fraction of a
micro second after the sweep generator is triggered (100m length of cable
may cause a delay of about 0.3 μs.
4.5Measurements of capacitance and loss tangent
Measurement of High Voltages J.R. Lucas 270
4.5.1 High Voltage Schering Bridge
High Voltage Schering Bridge is the method most widely used for
measuring capacitance and loss tangent (or power factor) in dielectrics.
Figure 4. - High Voltage
Schering Bridge
a.c.
sup
ply
Safety
gap
100 V
PDC
1
High
Voltage
Capacitor
High
Voltage
Standard
Capacitor
C
2
C
3
QSSafety
gap
100 V
Measurement of High Voltages J.R. Lucas 271
One arm is the high voltage test capacitor (represented by a series
combination of capacitance C
1
and resistance P for ease of analysis).
The other 3 arms are a standard high voltage capacitor C2 (generally a
loss free air capacitor of value 100 to 500 pF), a variable low resistance Q,
and a parallel combination of a standard low resistance S and a variable
capacitance C3.
The high voltage supply for the bridge is obtained through a high voltage
transformer.
For reasons of safety, only the high voltage test capacitor and the high
voltage standard capacitor will be at high voltage.
The other components are at low voltage and are not allowed to have
voltages greater than about 100 V applied across them by means of safety
gaps connected across them (The safety gaps are either gas discharge gaps
or paper gaps).
Measurement of High Voltages J.R. Lucas 272
The impedance of these arms must thus necessarily be of values much less
than that of the high voltage capacitors.
For measurements at power frequencies, the detector used is a vibration
galvanometer, usually of the moving magnet type (If the moving coil type
is used, it has to be tuned). The arms Q and C
3
are varied to obtain
balance.
δϕ
VI
Figure
4.
Measurement of High Voltages J.R. Lucas 273
It can be shown that this bridge is frequency independent, and that at
balance

C
C
=
Q
P
also
2
3
,
S
Q
=
C
C
1
2
power factor angle = φ,
loss angle = π/2 - φ = θ
θ ≈ tan θ = ω P C
1
= ω C
3
S
The low voltage end of the bridge is usually earthed, and since the
voltages across Q and S are limited to about 100 V, the detector would
also be near earth potential.
Thus all the variable arms and the detector can be safely handled by the
operator.
Measurement of High Voltages J.R. Lucas 274
It should be noted that the bridge is an unequal arm bridge, so that the
relative sensitivity will be small.
However, since the applied voltage is high, this is not a practical
disadvantage and a reasonable variation can be obtained across the
detector.
PDC
1
C
2
C
3
QSS

C
3

ab a.c.
supply
Measurement of High Voltages J.R. Lucas 275

Figure 4. - Bridge with guarded standard
capacitor
Measurement of High Voltages J.R. Lucas 276
Since the value of the standard capacitor must be accurately known, there
should be no distortion of the field in it.
Thus a high voltage guard is provided in it design. This guard is earthed
directly (which causes a small error), or kept at the same potential as the
main electrode without a direct connection as shown in figure 4.37.
The following procedure is used to have the guard electrode at the same
potential as the main electrode.
The bridge is adjusted for balance with the switch in position (a) - the
normal Schering Bridge.
Then with the switch in position (b) the bridge is again balanced using
only S' and C
3
'.
This ensures that finally a and b are at the same potential (same potential
as the other end of the detector).
Measurement of High Voltages J.R. Lucas 277
Successive balance is carried out in positions a and b alternately until
final balance is obtained.
This connection can be used for capacitances up to 2000 pF.
PDC
1
C
2
C
3
QSrpa.c.
supply
Measurement of High Voltages J.R. Lucas 278

Figure 4. - Schering Bridge for high capacitances
Measurement of High Voltages J.R. Lucas 279
When it is required to obtain higher value unknown capacitances (such as
in the case of a very long cable), the circuit is modified in the following
manner so that high current variable resistance standards would not be
required.
In this case we have a high current fixed value resistor shunted by a low
current variable high resistance which acts similar to a potential divider.
The expression at balance is obtained by converting the mesh consisting
of r, p and Q into star form, thus obtaining the normal schering bridge
arrangement.
At balance, it can be shown that
Measurement of High Voltages J.R. Lucas 280
]
]
]

p
C
r
C
- 1 S
C
= tan that so
3
2
3
ω θ
,
p
p + r + Q

Q
S

C
=
C
2 1
In the case of a cable already buried or earthed, then we would have to
earth the end of the source near the test capacitor.
Then all the equipment, and hence the operator would have to be at a high
voltage to earth.
The operator can either operate the instruments using long insulated rods,
or get into a Faraday cage (A cage which is raised to the same potential as
the high voltage electrode so that there is no difference in potential).
The earthing of the test capacitor near the detector end instead of the
source end would bring the instruments near earth potential, but is not
Measurement of High Voltages J.R. Lucas 281
used due to the introduction of stray capacitances by this means which
would cause measuring errors.
I
S
V
2
V
1
δδ
CC
S
s
YXV
1
V
2
I
S
Measurement of High Voltages J.R. Lucas 282
4.5.2 Dielectric loss measurement using Oscilloscope
Figure 4. - Dielectric loss measurement using oscilloscope C
s
- capacitance of sample
C - loss free capacitor (C >> C
s
)
δ - loss angle
Measurement of High Voltages J.R. Lucas 283
In an oscilloscope, if two alternating voltages of the same frequency are
applied to the x and y plates, the resulting figure will be an ellipse.
When the two voltages are in phase, the figure will be a straight line with
an enclosed area of zero.
As the phase angle increases, the area increases and reaches a maximum
when the phase angle difference is 90
o
.
This property is made use of in dielectric loss measurements.
A potential difference proportional to the applied voltage is applied to one
pair of plates and a potential difference proportional to the integral of the
current through the dielectric is applied to the other pair.
Since the loss is to be measured in a dielectric sample, a lossless large
capacitor is connected in series with the sample.
Voltages across capacitor and across sample applied across the 2 plates.
Measurement of High Voltages J.R. Lucas 284
The area of the ellipse thus formed is proportional to the power loss in the
dielectric. If the power loss in the dielectric is zero, the figure traced out
on the oscilloscope would be a straight line.
The use of the standard capacitor C ensures that the voltage across it is 90
o
out of phase with the current. Hence the angle on which the area of the
ellipse depends is not the power factor angle but the loss angle.
Power loss in C
s
= V
2
I
s
sin δ
on the oscilloscope
y-deflection ∝ v
1
= V
1m
sin (ω t - δ), and
x-deflection ∝ v
2
= V
2m
sin ω t, taken as reference.
i.e. y = a . V
1m
sin (ω t - δ) = a . (I
sm
/ω C) sin (ω t - δ)
and x = b . V
2m
sin ω t
where a, b are constants.
Measurement of High Voltages J.R. Lucas 285
The area of the ellipse traced out on the oscilloscope screen is given by

V I
.
2
.
C
b . a
=
dt . t . .
V
. b . ) - t ( .
C
I
. a = dx . y = A
2 s
2m
sm
T
0
δ
ω
π
ω ω δ ω
ω
sin
cos sin


It is thus seen that the area of the ellipse is proportional to the power loss.
Measurement of High Voltages J.R. Lucas 286
4.5.3 Detection of internal discharges
Detection of internal discharges can be carried out by various methods.
It can be done by
(a) visual methods - in transparent insulation the sparks can be detected by
either direct observation or by using a photo-electric cell;
(b) audible methods - the audible clicks given out by the discharges may
be detected by using a microphone, an ultrasonic detector or other
transducer; and
(c) electrical methods - these will be detailed out in the following
sections.
Measurement of High Voltages J.R. Lucas 287
Electrical Methods of discharge detection
(a) Using a corona detector
The discharge detector shown in the figure 4.40, is basically a wide band
amplifier with a gain of 10
6
and a bandwidth of 10 kHz to 150 kHz.
The lossy dielectric sample may be represented by a capacitance C
x
in
parallel with its discharge q
x
.
For the charge flow, it may be assumed that the high voltage supply
circuit provides almost infinite impedance, and that the step wave
generator has a negligible internal impedance.
Thus the discharge flow path is as shown on the diagram.
When the corona detector shows no discharge across it, the voltage drop
caused by the coupling capacitor C
q
must equal the voltage produced by
coupl
ing
capac
itor
coro
na
dete
ctor
step
wave
gener
ator
bloc
king
capa
citor
test
voltag
e
speci
men
disch
arge
Figure 4. - Discharge
detection using corona detector
C
q
C
b
C
x
q
x
e
q
qq
x
q
Measurement of High Voltages J.R. Lucas 288
the step wave generator, and the voltage across the blocking capacitor C
b
and by the specimen must must be zero.
Since the specimen has its own discharge in the opposite direction to q,
the total discharge through the specimen in the direction of q, must be q -
q
x
.
Since no drop occurs across the detector at balance,

C
q
=
C
1
+
C
1
q . e . i
0 =
C
/ ) q - (q +
C
q
also
C e
= q . e . i
x
x
x b
x
x
b
q q

,
`

.
|
0, =
C
q
-
e
q
q
Measurement of High Voltages J.R. Lucas 289

,
`

.
|
C
C
+ 1
C e
= q
b
x
q q
x
Substituting for q, we have
The energy dissipated in the void is given by
w = ½ q
x
V
0
where V
0
= peak voltage across specimen at inception voltage
(b) Using the oscilloscope with filtration and amplification
Internal discharges occurring within dielectric samples can be observed by
measuring the electrical pulses in the circuit where such discharges occur.
Measurement of High Voltages J.R. Lucas 290
choke dielectric
sample
loss-free
coupling
capacitor
k
h.v. transformer impedance Z high
pass
filter
h.f. amplifier CRO
Figure 4. - Circuit for discharge detection
Measurement of High Voltages J.R. Lucas 291
The apparatus used in the observation (namely the coupling capacitor and
the impedance) should be discharge tree, so that all the discharges caused
is due to the sample.
Measurement of High Voltages J.R. Lucas 292
low amplitude high frequency
pulses
superposed on sine wave
(power freq)
h.f.
pulses
Figure 4. - Output waveforms
Measurement of High Voltages J.R. Lucas 293
However, discharges occurring in the transformer and the choke are short
circuited through the coupling capacitor and do not affect the
measurement.
The discharge pulses caused in the sample are of high frequency, so that
we bypass the low frequency and amplify the high frequency in the
measurement circuit.
The coupling capacitor k is provided so that the high frequency
components would be provided with a low impedance path.
In the absence of this low impedance path, the path is highly inductive so
that these would act as high impedance to the high frequency.
(c) Using oscilloscope with elliptical time base
In many instances, the detector cannot be used close to equipment, and matching
units are employed which permit the use of about 30 m of co-axial lead between
detector and the source of discharge.
Measurement of High Voltages J.R. Lucas 294
Calibration is done by injection of a known step voltage into the system. This
gives direct calibration of discharge amplitude and takes into account the
response of the amplifier.
The discharge detector input circuit is shown in figure 4.43.
The output of the amplifier is displayed on a oscilloscope having an elliptical
time base.
The time base is produced from a phase shifting R-C network.
It is possible to distinguish between several types of discharges from the
nature of the output displayed on the oscilloscope.
co-axial
cable


Amplif
ier
Test
volta
ge
matching
unit

C
x
C
b
R R
m
R
m
Measurement of High Voltages J.R. Lucas 295
Displays on the oscilloscope for some typical
discharges are shown in figure 4.44 together with corresponding
waveforms arising out of external discharges as well as from contact
noise.
Figure 4. - Discharge
detector input circuit
P
o
s
i
t
i
v
e

p
e
a
k




m
a
r
k
e
r
P
o
s
i
t
i
v
e

p
e
a
k




m
a
r
k
e
r
P
o
s
i
t
i
v
e

p
e
a
k




m
a
r
k
e
r
P
o
s
i
t
i
v
e

p
e
a
k




m
a
r
k
e
r
n
e
g
a
t
i
v
e

p
e
a
k













m
a
r
k
e
r







































n
e
g
a
t
i
v
e

p
e
a
k













m
a
r
k
e
r







































negative peak

marker


negative peak

marker


p
u
l
s
e
C
a
l
i
b
r
a
t
i
o
n
p
u
l
s
e
C
a
l
i
b
r
a
t
i
o
n
p
u
l
s
e
C
a
l
i
b
r
a
t
i
o
n
p
u
l
s
e
C
a
l
i
b
r
a
t
i
o
n
Measurement of High Voltages J.R. Lucas 296

(
b
)













Figure 4.44 Displays for typical dicharges
(
a
)
(
c
)
















(
d
)
















n
e
g
a
t
i
v
e

p
e
a
k













m
a
r
k
e
r







































n
e
g
a
t
i
v
e

p
e
a
k













m
a
r
k
e
r







































Measurement of High Voltages J.R. Lucas 297
(a) For a typical oil-impregnated paper capacitor: The discharges are
approximately equal in magnitude and number in the two half cycles,
but have opposite polarity.
(b) For a polythene insulated cable: The discharges show the
asymmetry typical of discharges between a conductor and the solid
insulation for a polythene insulated cable.
(c) External discharges: Corona produces a very symmetrical display
about the negative voltage peak and as the voltage increases the
discharges spread over a larger part of the ellipse but remain
symmetrical.
(d) Contact noise: Bad contacts in the system produce many small
discharges at the current peaks.
(
b
)













(
c
)
















(
d
)
















Measurement of High Voltages J.R. Lucas 298
Oscilloscope connections for elliptical time base
The oscilloscope X and Y plates are supplied from a separate source so as
to form an ellipse on the screen.
By applying the output from the high frequency amplifier to the Y-plates,
we can obtain the high frequency pulse superposed on the ellipse.
The height of the pulse can be measured.
Knowing the voltage sensitivity of the scope, we can find the magnitude.
Knowing the characteristics of the amplifier we can calculate the output
from the circuit.
Then deriving a relation between the discharge from the sample and the
output across the impedance we can know the discharge from the sample.
Figure 4. - Generation of
elliptical time base
suppl
y for
ellipt
ical
time
base
Y
plat
es
X
plat
es
C
R
O
from
h.f.am
plifier
pulses
from
discha
rge
CR
Measurement of High Voltages J.R. Lucas 299
Calculation of internal discharge from measurements

void V
v
C
v
C
s
C
p
Measurement of High Voltages J.R. Lucas 300
e
q
u
i
v
al
e
n
t
ci
r
c
u
it
dielectric sample with void
F
i
g
u
r
e
4
.
-
E
q
u
i
v
a
l
e
n
t
c
i
r
c
u
it
o
f
d
i
e
l
e
c
t
r
i
c
w
it
h
v
o
i
d
Measurement of High Voltages J.R. Lucas 301
The internal discharges can be analysed by considering a single flaw in
the dielectric as shown in the figure 4.46.
Dielectric can be considered as being composed of a number of capacitances.
Between the two electrodes (other than in the strip containing the flaw), the
material is homogeneous and can be represented by a single capacitance
between the electrodes.
F
i
g
u
r
e
4
.
-
E
q
u
i
v
a
l
e
n
t
c
i
r
c
u
it
o
f
d
i
e
l
e
c
t
r
i
c
w
it
h
v
o
i
d
abcV
c
V
a
Figure 4.47 -
Equivalent
circuit
Measurement of High Voltages J.R. Lucas 302
The strip containing the flaw can also be considered as made up of three
capacitances in series; one representing the capacitance of the flaw and the other
two representing the capacitance on either side of the flaw.
The series capacitance on either side can be combined together to form a single
capacitance as shown in figure 4.47.
a - capacity of rest of dielectric
b - capacity of section of dielectric in series with cavity
c - capacity of cavity
F
i
g
u
r
e
4
.
-
E
q
u
i
v
a
l
e
n
t
c
i
r
c
u
it
o
f
d
i
e
l
e
c
t
r
i
c
w
it
h
v
o
i
d
Measurement of High Voltages J.R. Lucas 303
If the voltage across the cavity is greater than a certain critical value, then
the cavity would breakdown, the cavity capacitor discharges instantly, and
the voltage across the cavity would fall to zero.
V
c




-
Vc
Figure 4.48 - Discharge
waveforms across void
v







t












V
a
V
c

Critical
voltage
V
c

V
Measurement of High Voltages J.R. Lucas 304
Measurement of High Voltages J.R. Lucas 305
(a) (b)
wh
ere
Δu
-
me
asu
red
vol
tag
e
acr
oss
im
pe
da
nce
Z
ΔV
-
crit
ical
vol
tag
e
acr
oss
cav
ity
ΔV
1
-
Vo
lta
ge
acr
oss
sa
mp
le
Fi
gur
e
4.4
9 -
Cir
cui
t
for
an
aly
sis
a b
∆ V
a
∆ V
1
Ck
∆ u ∆ u∆ V
b Zk
Measurement of High Voltages J.R. Lucas 306
The cavity capacity again charges up within a very short period, and again
collapses discharging the charge.
This process repeats itself until the voltage across the cavity falls
below the critical value. This gives rise to a series of high frequency
pulses (each of duration of the order of 100 ns).
Figure 4.49 shows the actual circuit with the sample replaced by its
equivalent circuit.
Consider the case of the impedance Z being a capacitor C. The voltage
ΔV
1
would also be the voltage across the series combination of C and k.
Thus

wh
ere
Δu
-
me
asu
red
vol
tag
e
acr
oss
im
pe
da
nce
Z
ΔV
-
crit
ical
vol
tag
e
acr
oss
cav
ity
ΔV
1
-
Vo
lta
ge
acr
oss
sa
mp
le
Fi
gur
e
4.4
9 -
Cir
cui
t
for
an
aly
sis
Measurement of High Voltages J.R. Lucas 307
C + C/k) + (1 a) + (b
V . b
=
C + k
k
.
C + k
k C
+ a + b
b . V
= u
C + k
k
=
V
u
that so
1


∆ ∴




,
k + C
k C
+ a + b
b
=
V
V1
In this expression b.ΔV is the charge dissipated in the discharge.
Also, since the cavity is small, its capacity has negligible effect on the
total capacitance.
C + k) / C + (1 e) capacitanc (apparatus
q
= u ∆
If the impedance across which the voltage is measured is a parallel
combination of a capacitance C and a resistance R, then the above

wh
ere
Δu
-
me
asu
red
vol
tag
e
acr
oss
im
pe
da
nce
Z
ΔV
-
crit
ical
vol
tag
e
acr
oss
cav
ity
ΔV
1
-
Vo
lta
ge
acr
oss
sa
mp
le
Measurement of High Voltages J.R. Lucas 308
calculated value of voltage would correspond to the value before the
capacitor C discharges through the resistance R exponentially, and the
actual expression would be
C + k) / C + (1 e) capacitanc (apparatus
e
. q
= u
R C
t
-


wh
ere
Δu
-
me
asu
red
vol
tag
e
acr
oss
im
pe
da
nce
Z
ΔV
-
crit
ical
vol
tag
e
acr
oss
cav
ity
ΔV
1
-
Vo
lta
ge
acr
oss
sa
mp
le
with C only with C & parallel R
with C & parallel R & L
Δ u
Measurement of High Voltages J.R. Lucas 309
All the apparatus other than the sample should be as discharge free as
possible.
If there are external discharges other than from the sample, the value of
Δu would be due to the total discharge and the calculations would be in
error.
Figure 4.50 - Output
waveform
Measurement of High Voltages J.R. Lucas 310
A method of avoiding external discharges is by having a bridge type of
circuit as shown in figure 4.51.
Figure 4.51 - Circuit to avoid
effects of external discharges
external
discharges
sam
ple
Dischar
ges
from
sample

high pass
filteration,
amplificatio
n
and
detection
sup
ply
k
Measurement of High Voltages J.R. Lucas 311
In this circuit, external discharges would affect both the resistances
equally, so that if the detection is done across the two resistances, it would
measure only the discharges due to the internal flaws in the sample.
Measurement of High Voltages J.R. Lucas 312
4.5.4 Measurement of dielectric constant and dissipation factor of a
liquid dielectric at high frequencies using a resonance method
perspex
cover
brass
electr
ode
bras
s
test
cell
3m
m
gap
Figure 4.52
- Test Cell
Measurement of High Voltages J.R. Lucas 313
Test cell used in the measurement consists of a brass cell inside which is
suspended a brass electrode from a perspex cover.
The outer cell is the earthed electrode, and there is a gap of 3 mm all
round between this and the inner brass electrode.
Since electrodes are near each other, stray capacitances have to be
considered as well.
The test cell is connected in parallel with a variable capacitor and made
part of a resonant circuit as shown in figure 4.53.
In the circuit, R is a high series resistance used to keep the total current in
the circuit very nearly constant.
The stray capacitance C
0
of the test cell can be obtained by removing the
inner electrode of the test cell and with the empty cell resonance obtained.
Measurement of High Voltages J.R. Lucas 314
Measurement of High Voltages J.R. Lucas 315
R
C
v
a.c.
source
VV
coil
test
cell
L
r
Figure 4.53 - Test Circuit
C
v
o
L
r
C
o
Figure 4.54 - Equivalent
circuit for case (i)
Measurement of High Voltages J.R. Lucas 316
If C
v
is the value of the variable capacitor at resonance, at the angular
frequency ω, then
1 = )
C
+
C
( L
0 v
2
ω
The above calculation is required only if the stray capacitance value is
actually required.
Otherwise the stray capacitance can be eliminated using the following
procedure at the selected frequency (say 1 MHz).
Measurement of High Voltages J.R. Lucas 317
(i) With the outer cell and with only the brass screw and the perspex
cover of the inner cell in position, the variable capacitor C
v0
is varied until
resonance is obtained.
Under this condition, only the stray capacitance C
0
is present, and the total
capacitance will be at resonance with the coil inductance L.
The effective capacitance, in this case, is C
v0
+ C
0
.
The Q-factor of the circuit will be dependant on the resistance r of the
coil.
The test cell permits replacement of the 3 mm air gap by the insulating
liquid permitting comparison of capacitances to determine the dielectric
constant through the measurement of capacitance change at resonance.
The circuit is then de-tuned to the half-power points
V
m
V
m
/
√2
C
-

C
C
+
Measurement of High Voltages J.R. Lucas 318
If C+ and C– are the values at the half power points (voltage
corresponding to 1/√2), then it can be shown that the Q factor at
resonance can be obtained from
C
+
C
)
C
-
C
( + C 2
=
C
-
C
C
+
C
= Q
- +
- +
- +
- +
∆ ∆
∆ ∆
where C+ and ∆ C– are the variations at the half power points.
Usually Q is high, and
v
- +
, C =
C
=
C
C
C
= Q that so

∆ ∆ ∆
The Q-factor can be determined from the half-power points.
The variable capacitance is varied in either direction from resonance until
the half-power points (voltage corresponding to 1/√2) are reached.
Measurement of High Voltages J.R. Lucas 319
(ii) The inner electrode is now screwed in, and the circuit is again
adjusted for resonance at the same frequency.
C
v1
L
r
C
o
C
a
R
a
Figure 4.55 - Equivalent
circuit for case (ii)
Measurement of High Voltages J.R. Lucas 320
If C
a
is the capacitance of the active portion of the test cell with air as
dielectric, and R
a
is the equivalent shunt resistance of the circuit with air
as dielectric, then the total value of the capacitance required must remain
the same. This is true for all cases.
Thus we have


C
-
C
=
C

C
+
C
+
C
=
C
+
C
v v a
a 0 v 0 v
1 0
1 0

The Q-factor of the circuit however will be different from the earlier
value, due to the additional parallel resistance.
If the parallel equivalent resistance of the inductor is considered, then it is
seen that the overall Q factor Q
a
is given as the parallel equivalent of the
Q-factors of the coil resistance and the resistance R
a
.
The Q-factor corresponding to the resistance R
a
is ωCR
a
, so that
Measurement of High Voltages J.R. Lucas 321
R
C
1
+
Q
1
=
Q
1
a
L a
ω
(iii) The liquid is now introduced into the test cell.
Measurement of High Voltages J.R. Lucas 322

[The liquid level should be slightly below the perspex cover, so that the
surface condition of the perspex is not changed.]
If R
k
is the equivalent shunt resistance of the liquid, and k is the relative
permittivity of the liquid dielectric, then the capacitance of the active
portion of the test cell with the liquid would be kC
a
.
C
v2
L
r
C
o
C
k
R
a
R
k
Figure 4.56 - Equivalent
circuit for case (iii)
Measurement of High Voltages J.R. Lucas 323
If C
v2
is the value of the variable capacitor at resonance, then
C
-
C
C
-
C
= k
C
-
C
=
C
k giving
1 v 0 v
2 v 0 v
2 v 0 v a
2 0


C
k +
C
+
C
=
C
+
C
a 0 v 0 v
Also we have the equivalent Q factor Q
k
equivalent to the parallel
equivalent.
Thus
R
C
1
+
R
C
1
+
Q
1
=
Q
1
k a
L k
ω ω
Thus the inverse of ωCR
k
can be determined from
Measurement of High Voltages J.R. Lucas 324
C
) C (
=
Q
1
,
C
) C (
=
Q
1
using calculated be can
a
a
k
k
∆ ∆

Q
1
,
Q
1

Q
1
-
Q
1
=
R
C
1
a k
a k
k
ω
[ ]
[ ]
C
-
C
.
C
k
1
=
C
C
-
C
.
C
k
C
=
Q
1
-
Q
1
.
C
k
C
=


C
C
.
R
C
1
=
R C
1
= factor loss
a k
a
a k
a
a k
a
k k k k
∆ ∆
∆ ∆
]
]
]

ω ω
The loss factor of the dielectric is given by
Measurement of High Voltages J.R. Lucas 325
C
-
C
C
-
C
= factor loss . e . i
2 v 0 v
a k
∆ ∆
Note: in making connections it is essential that care is taken to minimise
stray capacitances by using short leads, and the components should not be
disturbed during the experiment.
hot
wire
h.v.
line
safety
gap
Geart
h
Measurement of High Voltages J.R. Lucas 326
4.5.5 Ionic Wind Voltmeter
Figure 4.57 - Ionic Wind
Voltmeter
Measurement of High Voltages J.R. Lucas 327
When a highly charged point is situated in air or other gas, a movement of
the air around the point is observed.
This is referred to as electric wind and is brought about by repulsion of
ions from surface of the point by the intense electro-static field.
These ions colliding with uncharged molecules of air carry them with it
setting up the electric wind.
A similar wind is observed also at the earth electrode.
In the ionic wind voltmeter, a hot wire, of platinum-gold alloy, included in
one arm of a Wheatstone bridge network is used as the earthed electrode
of high-tension gap.
Before the high voltage is applied, the bridge is balanced.
When the voltage of the gap exceeds the "threshold voltage" (voltage
required before the potential gradient is sufficient for ionization to
commence), the electric wind cools the hot wire and hence reduces the
resistance.
ellips
oid
shiel
d
Measurement of High Voltages J.R. Lucas 328
This reduction causes an appreciable out-of-balance voltage in the bridge.
The voltage waveform influences the instrument reading, and the
instrument is calibrated for a sine wave.
The voltmeter can be used to determine either the peak value or the r.m.s.
value of alternating voltages and direct voltages.
The principle advantages are that the h.v. may be measured by an observer
at some distance from the charged conductors, and the robust construction
and freedom from disturbances by temperature and weather conditions
which make it suitable for outdoor use.
4.5.6 Dumb-bell Voltmeter
F
i
g
u
r
e
4
.
5
8
-
D
u
m
b
-
b
e
ll
v
o
lt
m
e
t
e
r
Measurement of High Voltages J.R. Lucas 329
A rather specialized way of measuring r.m.s. voltage was developed by
F.M.Bruce in which the period of oscillation of a conducting spheroid in
an electro-static field was determined.
This enabled the voltage to be determined in terms of length and time,
with an accuracy of 0.05%.
The instrument is shown on figure 4.58.
F
i
g
u
r
e
4
.
5
8
-
D
u
m
b
-
b
e
ll
v
o
lt
m
e
t
e
r
5
High Voltage Generators
for Testing
F
i
g
u
r
e
4
.
5
8
-
D
u
m
b
-
b
e
ll
v
o
lt
m
e
t
e
r
331 J R Lucas
Generation of High Voltages for Testing
5.0Generation of High Voltages
Power systems engineers are interested in high voltages primarily for
power transmission, and secondly for testing of equipment used in power
transmission.
In this chapter, high voltages are generated for testing of insulation.
Generation has to be carried out in the testing laboratory.
In many testing laboratories, the primary source of power is at low voltage
(400 V three phase or 230 V single phase, at 50 Hz).
Since insulation is usually being tested, impedances involved are
extremely high (~ MΩ) and the currents small (<1 A).
Therefore high voltage testing does not usually require high power.
Thus special methods may be used which are not applicable when
generating high voltage in high power applications.
332 J R Lucas
Generation of High Voltages for Testing
5.1Generation of High Alternating Voltages
Single transformer test units are made for high alternating voltages up to
about 200 kV.
However, for high voltages to reduce the cost (insulation cost increases
rapidly with voltage) and make transportation easier, a cascade
arrangement of several transformers is used.
5.1.1 Cascade arrangement of transformers
Figure 5.1 shows a typical cascade arrangement of transformers used to
obtain up to 300 kV from three units each rated at 100 kV insulation.
The low voltage winding is connected to the primary of the first
transformer, and this is connected to the transformer tank which is
earthed.
One end of the high voltage winding is also earthed through the tank.
333 J R Lucas
Generation of High Voltages for Testing
The high voltage end and a tapping near this end is taken out at the top of
the transformer through a bushing, and forms the primary of the second
transformer.
334 J R Lucas
Insulati
ng
Pedesta
l
Insulati
ng
Pedesta
l
1
kV
99
kV
100
kV
1
kV
199
kV
200
kV
1
kV
200
kV
100
kV
hv
output
300
kV
bush
ing
E99
kV
100
kV
199
kV
200
kV
Generation of High Voltages for Testing

Figure 5. - Cascade arrangement of transformers
335 J R Lucas
Generation of High Voltages for Testing
One end of this winding is connected to the tank of the second
transformer to maintain the tank at high voltage.
The secondary of this transformer too has one end connected to the tank
and at the other end the next cascaded transformer is fed.
This cascade arrangement can be continued further if a still higher
voltage is required.
What is shown in the cascade transformer arrangement is the basic
principle involved.
The actual arrangement could be different for practical reasons.
I
n
t
h
e
c
a
s
c
a
d
e
a
r
r
a
n
g
e
m
e
n
t
s
h
o
w
n
,
e
336 J R Lucas
RCLEV
Generation of High Voltages for Testing
5.1.2 Resonant Transformers
The resonance principle of a series tuned L-C circuit can be made use of
to obtain a higher voltage with a given transformer.
Let R represent the equivalent parallel resistance across the coil and the
device under test. The current i would be given by
Figure 5. -
Resonance
circuit
337 J R Lucas
Generation of High Voltages for Testing
resonance at
L j
R . E
- =
R C L - L j + R
E . R C L -
= v . e . i

L j + R
R L j
. i = v that so
2
2
ω
ω
ω
ω
ω
ω
ω
ω
ω L j + R
R L j
+
C j
1
E
= i
Since R is usually very large, the Q factor of the circuit (Q = R/Lω)
would be very large, and the output voltage would be given by
Q . E =
L
R
. E = | v |
ω
It can thus be seen that a much larger value that the input can be obtained
across the device under test in the resonant principle.
338 J R Lucas
Generation of High Voltages for Testing
50 Hz
supply
CL
R
air-cored
coil
air-cored
transformer
Test
device
Figure 5. - resonant transformer
339 J R Lucas
Generation of High Voltages for Testing
C L
1
= f 2 =
resonance at
π ω
Figure 5.3 shows the application of the resonance principle at power
frequency.
For certain applications, particularly when the final requirement is a direct
voltage, it is an advantage to select a frequency higher than power
frequency (50 Hz).
This would result in a smaller transformer having fewer turns, and also
simplifies the smoothing after rectification.
High voltage high frequency voltages are not readily available, and the
following is sometimes used to obtain a supply at three times power
frequency.
340 J R Lucas
Generation of High Voltages for Testing
It makes use of the fact that the magnetising current of a transformer has a
high third harmonic component.
Thus if an open delta secondary is used, no power frequency voltage
would remain and only the third harmonic component would be present.
Figure 5.4 shows the circuit arrangement.
341 J R Lucas
Generation of High Voltages for Testing
Air-cored coils are used to simplify the construction and the insulation.
Figure 5. - Resonant transformer
with third harmonic
3-
phase
50
Hz
suppl
y
3
rd

harmon
ic
compo
nent
High
voltage
3
rd
-
harmonic
componen
t output
CL
342 J R Lucas
Generation of High Voltages for Testing
5.1.3 High frequency high voltages
High frequency (few kHz to Mhz) high voltages are required in testing
apparatus for behaviour with switching surges, insulation flashover etc.
The importance of testing with high frequency is that high frequency
oscillations cause failure of insulator at comparatively low voltage due to
high dielectric los and consequent heating.
Thus it is necessary to produce damped high frequency voltages.
The damped oscillations are obtained by the use of a Tesla coil, together
with a circuit containing a quenched spark gap.
The tesla coil constitutes the high voltage transformer.
It consists of two air-cored coils which are placed concentrically.
The high voltage secondary coil has a large number of turns, and is wound
on a frame of insulating material, the insulation between turns being air,
or in some cases, oil.
343 J R Lucas
Generation of High Voltages for Testing
The primary winding has only a few turns wound on an insulating frame.
The supply is usually 50 Hz to the primary of the high voltage testing
transformer.
[In the circuit shown, C
2
includes the capacitance of the sphere gap used
for measurement.]
The primary circuit of the tesla transformer also contains a trigger spark
gap.
Figure 5. - Tesla coil circuit for
high frequency generation
Measu
ring
sphere
gap
low
voltag
e
50 Hz
iron-
core
transfo
rmer
C
2
L
1
trig
ger
gap
L
2
equip
ment
under
test
Tesla
coil
C
1
344 J R Lucas
Generation of High Voltages for Testing
Since the supply to the primary of the tesla transformer is alternating, the
capacitor C
1
is charged up to some maximum voltage, which depends
upon the secondary side of the supply transformer, and upon the setting of
the trigger gap.
At this voltage, the trigger gap breaks down, the capacitor C
1
discharges,
and a train of damped oscillations of high frequency is produced in the
circuit containing C
1
, the spark gap and the primary winding of the tesla
transformer.
During the time taken for this train of oscillations to die away, the spark
gap is conducting, due to the formation of an arc across it.
This charge and discharge of capacitor C
1
takes place twice in one voltage
cycle.
Thus there will be a hundred of these trains of damped oscillations per
second.
345 J R Lucas
Generation of High Voltages for Testing
The frequency of oscillations themselves is very high (about 100 kHz
usually), its actual value depending upon the inductance and capacitance
of the oscillatory circuit.
The circuit parameters are generally such that the resonant frequencies of
the two sides are the same.
ω
2
2 2 1 1
1

L C
=
L C

The expression for the voltage variation being obtained as the solution to a
fourth order differential equation.
complex is whereα
α α α α
,
e
D +
e
C +
e
B +
e
A = v
t - t - t - t -
4 3 2 1
The solution to the differential equation will generally be in conjugate
pairs.
346 J R Lucas
Generation of High Voltages for Testing
etc .... ; j - a = , j + a =
2 1
ω
α
ω
α
Thus the solution can be written in the form
constants are , , , ,
A
,
A
,
a
,
a
where
sin sin
2 1
2 1 2 1 2 1
φ φ
ω ω
φ
ω
φ
ω
) + t (
e A
+ ) + t (
e A
= v
2
2
t
a
-
2
1
1
t
a
-
1
2 1
If the two undamped frequencies are equal
(corresponding to L
1
C
1
= L
1
C
1
),
then the damped resonant frequencies are nearly equal (ω
1
≈ ω
2
).
The exponential decays of the components of the voltage depends on the
resistance values.
347 J R Lucas
Generation of High Voltages for Testing
If amplitudes A
1
and A
2
are equal, and the decays also equal, then the
summation in v would have the form
2
) - + t . - (
.
2
) + + t . + (
2 = ) + t ( + ) + t (
2 1
2 1
2 1
2 1
2
2
1
1
φ φ
ω ω
φ φ
ω ω
φ
ω
φ
ω
cos sin sin sin
If ω
1
ω
2
, then (ω
1
+ ω
2
)/2 ≈ ω, so that the sum of the two sine terms
represents a product of terms, one of which is of very nearly the resonant
frequency, and the other with a frequency equal to the difference
frequency between the primary and the secondary resonance frequencies.
If the magnitudes and decays were not considered equal,
the above result will be modified by the constants A
1
and A
2
,
and the exponential decays e
-a
1
t
and e
-a
2
t
.
The energy tends to get transferred from primary to the secondary and
vice versa, so that the voltage of primary is minimum when the secondary
voltage is maximum and vice versa.
348 J R Lucas
Primary voltage Secondary voltage (a) Primary voltage (b)
F
i
g
u
r
e

5
.


-


V
o
l
t
a
g
e

w
a
v
e
f
o
r
m
s

a
c
r
o
s
s

t
e
s
l
a

t
r
a
n
s
f
o
r
m
e
r
Generation of High Voltages for Testing
Secondary voltage
349 J R Lucas
F
i
g
u
r
e

5
.


-


V
o
l
t
a
g
e

w
a
v
e
f
o
r
m
s

a
c
r
o
s
s

t
e
s
l
a

t
r
a
n
s
f
o
r
m
e
r
Generation of High Voltages for Testing
Oscillation would occur which would be damped out due to the resistance
in the circuit.
What we require is a single series of short duration pulses.
This can be done by preventing the energy from travelling backwards and
forwards in the tesla transformer by quenching the trigger gap by air blast
cooling.
When the primary voltage is zero, the blast of air removes the spark in the
primary gap so that the energy is confined to the secondary.
Figure 5.6 (a) shows the primary and secondary voltage waveforms
without quenching and figure 5.6 (b) shows the corresponding waveforms
with quenching.
350 J R Lucas
F
i
g
u
r
e

5
.


-


V
o
l
t
a
g
e

w
a
v
e
f
o
r
m
s

a
c
r
o
s
s

t
e
s
l
a

t
r
a
n
s
f
o
r
m
e
r
Generation of High Voltages for Testing
5.2Generation of High Direct Voltages
Generation of high direct voltages are required in the testing of high
voltage direct current apparatus as well as in testing the insulation of
cables and capacitors where the use of alternating voltage test sets become
impractical due to the steady high charging currents.
5.2.1 Rectifier circuits
One of the simplest methods of producing high direct voltages for testing
is to use either a half-wave for full-wave rectifier circuit with a high
alternating voltage source.
The rectifiers used must be high voltage rectifiers with a peak inverse
voltage of at least twice the peak value of the alternating voltage supply.
In theory, a low pass filter may be used to smooth the output, however
when the test device is highly capacitive no smoothing is required.
Even otherwise only a capacitance may be used across the test device for
smoothing.
351 J R Lucas
Generation of High Voltages for Testing
Figure 5.7 shows the half-wave and the full wave arrangements.
In testing with high voltage direct current care must be taken to discharge
any capacitors that may be present before changing connections.
In certain test sets, automatic discharging is provided which discharges
the capacitors to earth.
5.2.2 Voltage Multiplier Circuits
Both full-wave as well as half-wave circuits can produce a maximum
direct voltage corresponding to the peak value of the alternating voltage.
When higher voltages are required voltage multiplier circuits are used.
CCR
L
low
voltag
e
50 Hz
Rectif
ier
+–

R
L
+–

ABlow
voltag
e
50 Hz
(a) half-wave rectifier
(b) full-wave rectifier
Figure 5. - Half-wave and full-
wave rectifier circuits
352 J R Lucas
Generation of High Voltages for Testing
The common circuits are the voltage double circuit and the Cockroft-
Walton Circuit.
Voltage Doubler Circuit
The voltage doubler circuit makes use of the positive and the negative half
cycles to charge two different capacitors.
These are then connected in series aiding to obtain double the direct
voltage output.
Figure 5.8 shows a voltage doubler circuit.
In this case, the transformer will be of small rating that for the same direct
voltage rating with only simple rectification.
l.v.
sup
ply
Figure 5. - Voltage
doubler circuit
V
max

sin ωt
V
max
V
max
+ – 0
+ –
2V
max
353 J R Lucas
Generation of High Voltages for Testing
Further for the same direct voltage output the peak inverse voltage of the
diodes will be halved.
Cockroft-Walton Circuit
When more than doubling of the voltage is required, the Cockroft-Walton
voltage multiplier circuit is commonly used.
The circuit is shown in figure 5.9.
354 J R Lucas
Generation of High Voltages for Testing

Let V
max
be peak value of secondary voltage of high voltage transformer.
To analyze the behaviour, consider that charging of capacitors actually
takes place stage by stage rather than somewhat simultaneously.
EC
2
C
4
C
1
C
3
D
1
D
2
D
3
D
4
000aabbcd+HT V
max
a.c.
supply
Figure 5. - Cockroft-Walton Circuit
355 J R Lucas
Generation of High Voltages for Testing
This assumption will not invalidate the result but will make analysis easier
to follow.
Consider the first part of the circuit containing the diode D
1
, the capacitor
C
1
, and the secondary winding.
During the first negative half cycle of the applied voltage, the capacitor C
1
charge up to voltage V
max
.
Since during positive half cycle which follows, diode D
1
is reverse
biassed, capacitor C
1
will not discharge (or will not charge up in the other
direction) and the peak of this half cycle, the point a will be at 2 V
max
.
During the following cycles, the potential at a will vary between 0 and 2
V
max
, depending on whether the secondary voltage and the capacitor
voltage are opposing or assisting.
Initially, capacitor C
2
would be uncharged, and the voltage at b would be
zero.
356 J R Lucas
Generation of High Voltages for Testing
Thus as the voltage at a varies between 0 and 2 V
max
, the diode D
2
is
forward biassed, and the capacitor C
2
would charge to 2 V
max
.
Once the voltage at b has reached 2 V
max
, the voltage at a would be less
than or equal to the voltage at b.
Thus once C
2
has charged up, this diode too would be reverse biassed and
the capacitor C
2
would not discharge.
The voltage at b would now remain constant at 2 V
max
.
C
3
is also initially assumed uncharged.
Since the voltage at a varies between 0 and 2 V
max
, the diode D
3
would
initially be forward biassed for almost the whole cycle.
Capacitor C
3
charges until it reaches 2 V
max
when b is 2 V
max
and a is 0.
As the voltage at a again increases to 2 V
max
, the voltage at c increases,
and thus the diode D
3
is reverse biassed and C
3
would not discharge.
As a reaches 2V
max
voltage at c rises to 4V
max
, as C
3
has not discharged.
357 J R Lucas
Generation of High Voltages for Testing
Thus after charging up has taken place, the voltage at c varies between 2
V
max
and 4 V
max
.
Assuming C
4
also to be initially uncharged, since the voltage at b is a
constant at 2 V
max
and the voltage at c varies between 2 V
max
and 4 V
max
initially, during most of the cycle, the diode D
4
is forward biassed and C
4
charges up to the maximum difference between d and b (i.e. to 2 V
max
).
This occurs when the voltage at c is 4 V
max
and the voltage at d would
now be 4 V
max
.
As the voltage at c falls from 4 V
max
to 2 V
max
, since the capacitor C
4
has
charged up it would not discharge, since there is no discharge path.
Thus once the capacitors are charged up the voltage at d remains constant
at 4 V
max
.
This sequence of voltages gained is shown in Table 5.1.
358 J R Lucas
Generation of High Voltages for Testing
Cycle
Location
0

T/2
+
T

3T/2
+
2T

5T/2
+
3T

7T/2
+
4T

a
0 2Vm 0 2Vm 0 2Vm 0 2Vm 0
b
0 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm
c
0 0 2Vm 4Vm 2Vm 4Vm 2Vm 4Vm 2Vm
d
0 0 0 4Vm 4Vm 4Vm 4Vm 4Vm 4Vm
Table 1
When the generator is used for a test, or when it is loaded, a current is
drawn from the generator, and the capacitors lose some of their charge to
the load, and the voltage falls slightly depending on the load.
359 J R Lucas
Generation of High Voltages for Testing
As the voltage across any of the capacitors drops, then at some point in
the applied alternating voltage cycle, the corresponding diode would
become forward biassed and charging up of the capacitor would once
again result.
Thus when a load is connected, there would be a small ripple in the output
voltage.
5.2.3 Electrostatic generators
Electrostatic generators using the principle of charge transfer can give
very high direct voltages.
The basic principle involved is that the charge is placed on a carrier, either
insulating or an isolated conductor, and raised to the required potential by
being mechanically moved through the electrostatic field.
Van de Graeff generator
360 J R Lucas
++++++++++++++++++++++++++++++++++++++++++++++++++
Generation of High Voltages for Testing

Figure 5. - Van de Graeff Generator
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+


+ −
positive insulating
belt
moving
driver
pulley
corona spray
device
••••−−
−−
−−
−−
−−
−−
−−
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

+
positive +++

361 J R Lucas
Generation of High Voltages for Testing
The Van de Graeff generator is one of the
methods used to obtain very high
voltages.
However they cannot supply much
currents and the power output is restricted
to a few kilowatt, and their use is
restricted to low current applications.
The Van de Graeff generator uses an
insulating belt as the carrier of charge.
The generator consists of a low direct
voltage source, with corona discharge
taking place at the positive end of the
source.
The corona formation (spray) is caused
by a core like structure with sharp points (corona spray device).
362 J R Lucas
Generation of High Voltages for Testing
Charge is sprayed onto the belt at the bottom by corona discharges at a
potential of 10 to 100 kV above earth and carried to the top of the column
and deposited at a collector.
The upper electrode at which the charge is collected has a high radius of
curvature and the edges should be curved so as to have no loss.
The generator is usually enclosed in an earthed metallic cylindrical vessel
and is operated under pressure or in vacuum.
The higher voltage of the upper electrode arises from the fact that for the
same charge, a smaller capacitance gives a larger voltage.
The upper electrode has a smaller capacitance to earth on account of the
larger spacing involved.
C
Q
= V
363 J R Lucas
Generation of High Voltages for Testing
The potential of the high voltage electrode rises at a rate of
current charging net the is I where
C
I
=
t d
Q d

C
1
=
t d
V d
A steady potential will be reached by the high voltage electrode when the
leakage currents na the load current are equal to the charging current.
The edges of the upper electrode are so rounded as to avoid corona and
other local discharges.
With a single source at the lower end, the belt moves upwards with a
positive charge and returns uncharged.
364 J R Lucas
Generation of High Voltages for Testing
Charging can be made more effective by having an additional charge of
opposite polarity sprayed onto the belt by a self inducing arrangement
(negative corona spray) using an ingenious method.
This arrangement effectively doubles the charging rate.
365 J R Lucas
- - + +
- +
- +

- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- - + +

H.T. Exciter Rotor
Generation of High Voltages for Testing
Sames Generator
F
i
g
u
r
e
5
.
-
S
a
m
e
s
G
e
n
e
r
a
t
o
r
366 J R Lucas
- - + +
- +
- +

- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- +
- - + +

Generation of High Voltages for Testing
This is a more recent form of the electrostatic generator.
In this, the charge is carried on the surface of an insulating cylinder.
A 2 pole generator is shown in figure 5.11, but other number of poles are
also possible.
In this the power output will depend on the size of rotor.
The number of poles will determine the current and the voltage.
For example, a four pole rotor will produce twice the current at half the
voltage of that of a two pole machine of the same size.
In the Sames generator, the rotor is a hollow cylinder made of an
insulating material.
Electric charges are deposited on the surface of the rotor which is driven
by an electric motor to effect the transfer of charges in the field.
The whole unit is sealed in a pressure unit and insulated ith hydrogen at a
pressure of 10 to 25 atmospheres.
F
i
g
u
r
e
5
.
-
S
a
m
e
s
G
e
n
e
r
a
t
o
r
367 J R Lucas
Generation of High Voltages for Testing
6
High Voltage Testing
369 J R Lucas
High Voltage Testing
6.0High Voltage Testing Procedure
Electrical equipment must be capable of withstanding overvoltages during
operation.
Thus by suitable testing procedure we must ensure that this is done.
High voltage testing can be broadly classified into testing of insulating
materials (samples of dielectrics) and tests on completed equipment.
The tests carried out on samples of dielectric consist generally of the
measurement of permittivity, dielectric loss per unit volume, and the
dielectric strength of the material.
The first two can be measured using the High Voltage Schering Bridge.
The tests carried out on completed equipment are the measurement of
capacitance, the power factor or the total dielectric loss, the ultimate
breakdown voltage and the flash-over voltage.
370 J R Lucas
High Voltage Testing
The breakdown voltage tests on completed equipment is only done on a
few samples since it permanently damages and destroys the equipment
from further use.
However since all equipment have to stand up to a certain voltage without
damage under operating conditions, all equipment are subjected to
withstand tests on which the voltage applied is about twice the normal
voltage, but which is less than the breakdown voltage.
371 J R Lucas
High Voltage Testing
6.1General tests carried out on High voltage equipment
6.1.1 Sustained low-frequency tests
Sustained low frequency tests are done at power frequency (50 Hz), and
are the commonest of all tests.
These tests are made upon specimens of insulation materials for the
determination of dielectric strength and dielectric loss, for routine testing
of supply mains, and for work tests on high voltage transformers,
porcelain insulators and other apparatus.
Since the dielectric loss is sensitive to electric stress, the tests are carried
out at the highest ultimate stress possible.
For testing of porcelain insulators and in high tension cables, voltages as
high as 2000 kV may be used.
372 J R Lucas
High Voltage Testing
High voltage a.c. tests at 50 Hz are carried out as Routine tests on low
voltage (230 or 400 V) equipment.
Each one of these devices are subjected to a high voltage of 1 kV + 2 ×
(working voltage).
A 230 V piece of equipment may thus be subjected to about 1.5 to 2 kV.
These tests are generally carried out after manufacture before installation.
The high voltage is applied across the device under test by means of a
transformer. The transformer need not have a high power rating.
If a very high voltage is required, the transformer is usually build up in
stages by cascading. By means of cascading, the size of the transformer
and the insulation bushing necessary may be reduced in size.
373 J R Lucas
High Voltage Testing
The transformers are usually designed to have poor regulation so that if
the device under test is faulty and breakdown occurs, the terminal voltage
would drop due to the high current caused.
A resistance of about 1 ohm/volt is used in series with the transformer so
as to limit the current in the event of a breakdown to about 1 A.
The resistance used could be of electrolyte type (which would be far from
constant, but would be a simple device) such as a tube filled with water.
1
Ω /
V
200
kΩ
Dev
ice
und
er
test
200
kV
Figure 6. - a.c.
generation test circuit
374 J R Lucas
High Voltage Testing
In all high voltage tests, safety precautions are taken so as to ensure that
there is no access to the testing area when the high voltage is on.
There would be switches that would automatically be operated when the
door to the area is opened etc.
375 J R Lucas
High Voltage Testing
6.1.2 High Voltage direct current tests
These tests are done on apparatus expected to operate under direct voltage
conditions, and also where, due to the inconvenience of the use of high
capacity transformers required for extra high tension alternating voltage
tests and due to transport difficulties, alternating voltage tests cannot be
performed after installation.
A special feature of importance of the d.c. test is the testing of cables
which are expected to operate under a.c. conditions.
If the tests are done under a.c. conditions, a high charging current would
be drawn and the transformer used would have to have a current rating.
It is thus normal to subject the cable (soon after laying it, but before
energising it ) to carry out a high voltage test under d.c. conditions.
The test voltage would be about 2 (working voltage ) and the voltage is
maintained from 15 min to 1.5 hrs.
376 J R Lucas
High Voltage Testing
This d.c test is not complete equivalent to the corresponding a.c.
conditions , it is the leakage resistance which would determine the voltage
distribution, while in the a.c. conditions, it is the layers of different
dielectrics that determine the voltage distribution in the cable.
Although the electric field differs in the 2 cases, it is likely that the cable
will stand up to the required a.c. voltage.
The methods used to generate these high d.c. voltages have already been
described.
377 J R Lucas
High Voltage Testing
6.1.3 High-frequency tests
High frequency tests at frequencies varying from several kHz are
important where there is a possibility of high voltage in the lines etc., and
in insulators which are expected to carry high frequency such as radio
transmitting stations.
Also in the case of porcelain insulators, breakdown or flashover occurs in
most cases as a result of high frequency disturbances in the line, these
being due to either switching operations or external causes.
It is also found that high frequency oscillations cause failure of insulation
at a comparatively low voltage due to high dielectric loss and heating.
High voltage tests at high frequency are made at the manufacturing works
so as to obtain a design of insulator which will satisfactorily withstand all
conditions of service.
378 J R Lucas
High Voltage Testing
F
i
g
u
r
e

6
.
M
o
v
e
m
e
n
t
o
f
c
h
a
r
g
e
c
a
rr
i
e
r
s
379 J R Lucas
High Voltage Testing
In the case of power line suspension insulators, it is possible that
breakdown or flash over would occur due to high frequency over voltages
produced by faults or switching operations in the line.
Sudden interruptions in the line would give rise to resonant effects in the
line which would give rise to voltage waves in the line of high frequency.
These might cause flashover of the insulators.
The behaviour of insulating materials at high frequencies are quite
different to that at ordinary power frequency.
The dielectric loss per cycle is very nearly constant so that at high
frequencies the dielectric loss is much higher and the higher loss causes
heating effects.
The movements of charge carriers would be different.
At high frequency the polarity of electrodes might have changed before
the charge carriers have travelled from one electrode to the other, so that
they may go about half–way and turn back (figure 6.2).
F
i
g
u
r
e

6
.
M
o
v
e
m
e
n
t
o
f
c
h
a
r
g
e
c
a
rr
i
e
r
s
380 J R Lucas
High Voltage Testing
There are two kind of high frequency tests carried out.
(a) Tests with apparatus which produces undamped high–frequency
oscillations.
Undamped oscillations do not occur in power systems, but are useful for
insulation testing purposes especially for insulation to be in radio work.
(b) Tests with apparatus producing damped high–frequency oscillations.
When faults to earth or sudden switching of transmission lines occur, high
frequency transients occur whose frequency depends on the capacitance
and inductance of the line and will be about 50 kHZ to about 200 kHZ.
These are damped out with time.
6.1.4 Surge or impulse tests
These tests are carried out in order to investigate the influence of surges in
transmission lines, breakdown of insulators and of the end turns of
transformer connections to line.
381 J R Lucas
High Voltage Testing
In impulse testing, to represent surges generated due to lightning, the IEC
Standard impulse wave of 1.2/50 μs wave is generally used.
By the use of spark gaps, conditions occurring on the flash over to line are
simulated.
The total duration of a single lightning strike os about 100 μs, although
the total duration of the lightning stroke may be a few seconds.
Overvoltages of much higher duration also arise due to line faults,
switching operations etc, for which impulse waves such as 100/5000 μs
duration may be used.
In surge tests it is required to apply to the circuit or apparatus under test, a
high direct voltage whose value rises from zero to maximum in a very
short time and dies away again comparatively slowly.
Methods of generating such voltages have already been discussed earlier.
382 J R Lucas
High Voltage Testing
While impulse and high frequency tests are carried out by manufacturers ,
in order to ensure that their finished products will give satisfactory
performance in service , the most general tests upon insulating materials
are carried out at power frequencies .
383 J R Lucas
High Voltage Testing
Flash–over Tests
Porcelain insulators are designed so that spark over occurs at a lower voltage
than puncture, thus safeguarding the insulator, in service against destruction
in the case of line disturbances.
Flash–over tests are very importance in this case .
The flash–over is due to a breakdown of air at the insulator surface, and is
independent of the material of the insulator.
As the flash–over under wet conditions and dry conditions differ , tests such
as the one minute dry flash–over test and the one minute wet flash–over test
are performance.
(i) 50 percent dry impulse flash–over test, using an impulse generator
delivering a positive 1/50 μs impulse wave.
The voltage shall be increased to the 50 percent impulse flash–over voltage
(the voltage at which approximately half of the impulses applied cause flash–
over of the insulator)
384 J R Lucas
High Voltage Testing
(ii) Dry flash–over and dry one–minute test
In this the test voltage (given in the B.S.S.) is applied .
The voltage is raised to this value in approximately 10 seconds and shall
be maintained for one minute.
The voltage shall then be increased gradually until flash– over occurs .
(iii) Wet flash–over and one minute rain test
In this the insulator is sprayed throughout the test with artificial rain
drawn from source of supply at a temperature within 10 degrees of
centigrade of the ambient temperature in the neighborhood of the
insulator.
The resistivity of the water is to be between 9,000 and 11,000 ohm cm.
In the case of the testing of insulating materials , it is not the voltage
which produces spark–over breakdown which is important , but rather the
voltage for puncture of a given thickness ( ie. dielectric strength ).
385 J R Lucas
High Voltage Testing
The measurements made on insulating materials are usually, those of
dielectric strength, dielectric loss and power factor, the latter been
intimately connected with the dielectric strength of the material.
The dielectric strength of a given material depends, apart from chemical
and physical properties of the material itself, upon many factors including,
(a) thickness of the sample tested
(b) shape of the sample
(c) previous electrical and thermal treatment of the sample
(d) shape , size , material and arrangement of the electrodes
(e) nature of the contact which the electrodes make with the sample
(f) waveform and frequency of the applied voltage (if alternating )
(g) rate of application of the testing voltage and the time during which it is
maintained at a constant value .
(h) temperature and humidity when the test is carried out
(i) moisture content of the sample.
386 J R Lucas
High Voltage Testing
6.2Testing of solid dielectric materials
6.2.1 Nature of dielectric breakdown
Dielectric losses occur in insulating materials, when an electrostatic field
is applied to them.
These losses result in the formation of heat within the material.
Most insulating materials are bad thermal conductors, so that even though
heat so produced is small, it is not rapidly carried away by the material.
Now, the conductivity of such materials increases considerably with
increase of temperature, and the dielectric losses, therefore, rise and
produce more heat, the temperature thus building up from the small initial
temperature rise.
If the rate of increase of heat dissipated, with rise of temperature, is
greater than the rate of increase of dielectric loss with temperature rise, a
stable condition (thermal balance) will be reached.
387 J R Lucas
High Voltage Testing
If the latter rate of increase is greater than the former, the insulation will
breakdown owing to the excessive heat production which burns material.
The dielectric losses per cubic centimetre in a given material and at a
given temperature, are directly proportional to the frequency of the
electric field and to the square of the field strength.
Hence the decrease in breakdown voltage with increasing time of
application, increasing temperature and also the dependence voltage upon
shape, size and material of electrodes and upon the form the electric field.
The measurement of dielectric loss in insulating materials are very
important, as they give a fair indication as to comparative dielectric
strengths of such materials.
In the case of cable, dielectric loss measurements are now generally
recognized as the most reliable guide to the quality and condition of the
cable.
388 J R Lucas
High Voltage Testing
F
i
g
u
r
e

6
.
-
s
o
li
d
s
a
m
p
l
e
spherical
electrode
d
ie
le
ct
ri
c
m
at
e
ri
al
389 J R Lucas
High Voltage Testing
6.2.2 Determination of dielectric strength of solid dielectrics
A sheet or disc of the material of not less than 10cm in diameter, is taken
and recessed on both sides so as to accommodate the spherical electrodes
(2.5 cm in diameter) with a wall or partition of the material between them
0.5mm thick.
The electrical stress is applied to the specimen by means of the two
spheres fitting into the recesses without leaving any clearance, especially
at the centre.
The applied voltage is of approximately sine waveform at 50Hz.
This voltage is commenced at about 1/3 the full value and increased
rapidly to the full testing voltage.
Sometimes insulators after manufacture are found to contain flaws in the
form of voids or air spots.
F
i
g
u
r
e

6
.
-
s
o
li
d
s
a
m
p
l
e
390 J R Lucas
High Voltage Testing
These spots (due to non-homogeneity) have a lower breakdown strength
than the material itself, and if present would gradually deteriorate and
cause ultimate breakdown after a number of years.
High degree ionisations caused in these spots would give rise to high
energy electrons which would bombard the rest of the material, causing
physical decomposition.
In plastic type of materials,there might be carbonizations,
polymerisations, chemical decomposition etc.,which would gradually
diffuse into the material the by-products, causing chemical destruction.
The useful life of a component using such material will depend on the
weak spots and the applied voltage.
If the applied voltage is small, the life of the component is longer.
From design considerations the voltage to be applied if a particular life
span is required can be calculated.
391 J R Lucas
High Voltage Testing
F
i
g
u
r
e

6
.
-
v
a
ri
a
ti
o
n
o
f
l
o
s
s
f
a
c
t
o
r
a
p
p
li
e
d
v
o
lt
a
g
e
m
a
x
i
m
u
m

s
a
f
e
w
o
r
k
i
n
g
v
o
lt
a
g
e
l
o
s
s
f
a
c
t
o
r


t
a
n
δ
392 J R Lucas
High Voltage Testing
The schering bridge type of measurement gives an average type of
measurement,where the p.f. and the power loss indicates the value over
the whole of the length.
Thus small flaws if present would not cause much of a variation in the
overall p.f.
In the schering bridge type of measurement such flaws would not be
brought out.
The loss factor of a material does not vary much for low voltages.
As the voltage is increased at a certain value it starts increasing at a faster
rate.
This is the long time safe working voltage, since beyond this, the
specimen would keep on deteriorating.
If the apparatus need be used only for a short period, the applied voltage
can be higher than this safe value.
F
i
g
u
r
e

6
.
-
v
a
ri
a
ti
o
n
o
f
l
o
s
s
f
a
c
t
o
r
m
a
x
i
m
u
m

s
a
f
e
w
o
r
k
i
n
g
v
o
lt
a
g
e
l
o
s
s
f
a
c
t
o
r


t
a
n
δ
393 J R Lucas
High Voltage Testing
In a long length of
cable, the greater part of the cable would be in good conditions but with a
few weak spots here and there.
In a Schering bridge type of measurement, since it measures the overall
loss, such small individual spots cannot be detected.
It is necessary that such spots are detected as these increase with time and
finally cause its breakdown.
In high voltage transformers also there might be such small discharges
occurring which would not be measured by the schering bridge.
The method is to apply suitable high voltage to sample, and subject it to a
number of duty cycles ( heat cycles, make and break cycles).
Discharges caused are made to give pulses to a high frequency amplifier.
Figure 6. - sample of cable showing weak spots
dielectric conductor weak spots
394 J R Lucas
High Voltage Testing
The discharges caused are observed before and after such duty cycles to
see whether there is any appreciable increase in the pulse intensity after
the cycle of operation.
The methods of discussion have been discussed in an earlier chapter.
6.3Impulse Testing
395 J R Lucas
vt(i
)
vt(i
i)
vt(ii
i)
High Voltage Testing
Figure 6. - Observed impulse waveforms
396 J R Lucas
High Voltage Testing
These are done as tests on sample of apparatus.
The impulse test level is determined by the operating level (4 to 5 times
the normal operating value )
Apply on to the sample a certain number (say 10) positive impulse and 10
negative impulses of this particular value.
They should withstand this voltage without any destruction.
To test the ultimate impulse strength, apply increasing amounts of impulse
voltage until destruction occurs; during the tests it is necessary to see
whether there is any damage.
The damage may not be immediately visible, so we have it on a high
frequency ( single sweep and high speed ) oscilloscope.
In the event of complete damage, breakdown of the insulator due to the
application of the impulse voltage will be indicated as in (i).
397 J R Lucas
High Voltage Testing
If the insulator has suffered only a minor damage the wave form would
show no distortion , but would show as in (ii).
If there is no damage caused due to the impulse, the waveform will be
complete and undistorted as in (iii).
In testing high voltage insulators whose actual breakdown is in air (i.e
flashover takes place before breakdown of insulator), the porcelain itself
can be tested by immersing whole insulator in liquid of high permeability.
Thus there would be no outside flashover, and actual breakdown of the
insulator would occur.
398 J R Lucas
liq
uid
punct
ure
exter
nal
flash
over
High Voltage Testing
In specifying the
flashover characteristic in air we give the 50% flashover characteristic.
This is done as flashover occur at the same voltage on each application of
the impulse .
F
i
g
u
r
e

6
.
-
B
r
e
a
k
d
o
w
n
o
f
i
n
s
u
l
a
t
o
r
u
n
it
399 J R Lucas
High Voltage Testing
We apply different values of test voltages (impulse) and the voltage at
which there is 50% probability of breakdown is taken as 50% flashover
voltage.
F
i
g
u
r
e

6
.
-
B
r
e
a
k
d
o
w
n
o
f
i
n
s
u
l
a
t
o
r
u
n
it
F
i
g
u
r
e

6
.
-
P
r
o
b
a
b
il
it
y
o
f
fl
a
s
h
o
v
e
r
I
m
p
u
l
s
e
t
e
s
t
v
o
lt
a
g
e
5
0
%

b
r
e
a
k
d
o
w
n
v
o
lt
a
g
e
5
0
%
1
0
0
%
P
r
o
b
a
b
il
it
y

o
f
fl
a
s
h
o
v
e
r
400 J R Lucas
High Voltage Testing
The impulse flashover voltage also depends on the time lag of the applied
impulse before flashover time lag of the applied impulse before flashover
occurs.
Thus we have also got to determine the time lag characteristics for
breakdown.
If the voltage remains above a critical value long enough, flashover
occurs.
The time lag before flashover occurs depends on the statistical time lag
and on the formation time lag.
Depending on the volume of space between the gap, and also depending
on the nature of shielding, a certain time will be taken for enough free
electrons to be set free.
This is the statistical time lag.
F
i
g
u
r
e

6
.
-
P
r
o
b
a
b
il
it
y
o
f
fl
a
s
h
o
v
e
r
F
i
g
u
r
e

6
.
-
C
h
o
p
p
e
d
i
m
p
u
ls
e
w
a
v
e
f
o
r
m
ti
m
e
fl
a
s
h
o
v
e
r
I
m
p
u
l
s
e












v
o
lt
a
g
e
401 J R Lucas
High Voltage Testing
Once the electrons appear, depending on the voltage applied, they
multiply and ionise the space. once the space becomes conducting,
flashover occurs.
This is formation time lag.
To determine the time lag characteristic of a device, we can use the
impulse generator to generate impulses of gradually increasing amplitude
and determine the time of breakdown.
At each value, the test must be repeated a number of times so as to obtain
consistent values.
F
i
g
u
r
e

6
.
-
C
h
o
p
p
e
d
i
m
p
u
ls
e
w
a
v
e
f
o
r
m
I
m
p
u
l
s
e












v
o
lt
a
g
e
402 J R Lucas
High Voltage Testing
This type of characteristic is important in designing insulators.If a rod gap
is to protect a transformer.
Then the breakdown voltage characteristic of the rod gap must be less
than that of the transformer so as to protect it.
If the characteristic cross, protection will be offered only in the region
where the rod gap characteristic is lower than that of the transformer.
F
i
g
u
r
e

6
.
-
C
h
o
p
p
e
d
i
m
p
u
ls
e
w
a
v
e
f
o
r
m
I
m
p
u
l
s
e












v
o
lt
a
g
e
Figure 6. - Time lag
characteristic
tVolta
ge
403 J R Lucas
High Voltage Testing
System Voltage
I.E.C.
Impulse Withstand Voltage
11 kV 75 kV
33 kV 170 kV
66 kV 325 kV
132 kV 550 kV
275 kV 1050 kV
In obtaining the breakdown characteristic of a transformer we do not
attempt such tests that cause total destruction on transformers as they are
expensive.
What is done is we take a sample of the material used as insulators for the
transformers and then apply these tests till puncture takes place.
Thus the transformer characteristic is obtained by such tests on samples.
404 J R Lucas
High Voltage Testing
To obtain one point on the voltage vs time lag characteristic we would have
to do a large number of tests and take the mean,as these values vary from
sample to sample.
The sample would have to be surrounded by a liquid material of high
permittivity so that external flashover would not occur.
The impulse test voltage recommended by I.E.C. (International
Electrotechnical Commission) are given in the table 6.1.
The recommendation is that device when subjected to this voltage should not
suffer permanent damage or minor partial damage.
The voltage is set at slightly less than the withstand voltage and gradually
increase to test value.
About 10 positive impulses and 10 negative impulses are applied.
6.4Voltage Distribution in a Transformer Winding
Consider the entering of an impulse voltage on the terminating
transformer, as shown in figure 6.11.
405 J R Lucas
lightni
ng
arresto
r
transformer
winding
voltag
e
ti
m
e
surge
High Voltage Testing
Due to the presence of the interwinding capacitance and the capacitances
to earth of the transformer windings, the upper elements of the
transformer windings tend to be more heavily stressed than the lower
portions.
Due to the velocity of propagation of the impulse voltage would not be
evenly distributed in the winding.
F
i
g
u
r
e

6
.
-
S
u
r
g
e
p
r
o
p
a
g
a
ti
o
n
i
n
tr
a
n
s
f
o
r
m
e
r
w
i
n
d
i
n
g
406 J R Lucas
High Voltage Testing
Due to sharp rise of the voltage of the surge, there is a large difference of
voltage caused in the winding as the wave front travels up the winding.
Thus there would be an overvoltage across adjacent windings.
Depending on the termination, there will be reflections at the far end of
the winding.
If the termination is a short circuit, at the lowest point the voltage wave
whose amplitude is same as the original wave but of opposite polarity is
reflected.
For a line which is open circuited, the reflected wave would be of the
same magnitude and of the same sign.
Arising out of the reflections at the far end , there would be some coils
heavily stressed.
The position of the heavily stressed coils depending on the velocity of
propagation.
F
i
g
u
r
e

6
.
-
S
u
r
g
e
p
r
o
p
a
g
a
ti
o
n
i
n
tr
a
n
s
f
o
r
m
e
r
w
i
n
d
i
n
g
407 J R Lucas
High Voltage Testing
If flashover occurs at the gap (lightning arrestor) the voltage of the
impulse suddenly drops to zero when flashover occurs.
This can be represented by a full wave, and a negative wave starting from
the time flashover occurs.
The chopped wave, though it reduces the voltage of the surge to zero, will
have a severe effect of the winding due to sharp drop in the voltage.
Thus it is always necessary to subject the transformer during tests to
chopped wave conditions.
Generally the method is to apply full-waves and see whether damage has
occurred and then to apply the chopped waves and to see whether damage
has occurred and then to apply the chopped waves and to see whether
damage has occurred.
Example
F
i
g
u
r
e

6
.
-
S
u
r
g
e
p
r
o
p
a
g
a
ti
o
n
i
n
tr
a
n
s
f
o
r
m
e
r
w
i
n
d
i
n
g
408 J R Lucas
High Voltage Testing
A rectangular voltage V is impressed at the line terminal of a winding of a
high voltage transformer , the neutral point being isolated from earth.The
capacitance to earth of the complete winding is C
w
.
Prove that the voltage at a point in the winding distant X from the neutral
is
winding of length = l ,
C
C
=
a
where
cosh
cosh
w
g
2
,
a
x
l
a

. V =
v
x
If C
g
= 900 pF and Cw

= 10 pF , calculate the ratio of the maximum
initial voltage gradient in the winding to the average voltage gradient.
How will the initial voltage distribution in the winding be effected if the
wavefront has a duration of several microseconds ?
409 J R Lucas
line surge C
g
C
w
eart
h
Nv
q
d
x
x d
x
x
High Voltage Testing
In the case of a voltage with a vertical front incident on the transformer
winding , the voltage variation being instantaneous , the charging up is
instantaneous and the presence of the inductance of the winding may be
neglected.
F
i
g
u
r
e

6
.
410 J R Lucas
High Voltage Testing
0 = v .
l
1
.
C
C
-
x
v

q =
x
v
. l .
C
, v .
l
C
=
x
q

q = l .
x d
C
. x d .
x
v
, v . x d .
l
C
= x d .
x
q
Thus,
2
w
g
2
2
w
g
w g












x
l
a
sinh B + x
l
a
cosh A = v that so V; = v l, = x at
sinh cosh x
l
a
B + x
l
a
A = v ∴
411 J R Lucas
High Voltage Testing
. initially
a cosh
x .
l
a
sinh .
l
a
. V
=
x
v
also ,
a cosh
x .
l
a
cosh V
= v
a cosh
V
= A giving , 0 = B
0 = 0) cosh
l
a
. B + 0 sinh
l
a
. (A l
C
=
x
v
. l
C
= q that so , 0 = q , 0 = x at
w w






initially 9.48 a = stress rage stress/ave maximum

l
V

a cosh
1
- 1 .
l
V
=
x d . x .
l
a
sinh .
a cosh . l
a . V

l
1
= stress average

l
a
. V (9.48) tanh .
l
a
. V = l) = x (at gradient voltage initial Maximum
9.48 = a , 90 =
10
900
=
a
have we figures ng Substituti
l
o
2
≈ ∴

]
]
]




412 J R Lucas
1
x/l
1
x/l
(neutra
l) 0
V
ξ

x
High Voltage Testing
The voltage
F
i
g
u
r
e

6
.
-
v
a
ri
a
ti
o
n
s
o
f
v
o
lt
a
g
e
a
n
d
v
o
lt
a
g
e
g
r
a
d
i
e
n
t
413 J R Lucas
High Voltage Testing
distribution along the winding and the stress distribution (initially) are as
shown in figure 6.13.
If the wave-front time is of several micro-seconds duration, the charging
up would not be instantaneous, and the effect of the inductance during this
period may not be neglected.
The winding then behaves similar to a transmission line with distributed
inductances and shunt capacitances.
The effect of the surge would cause a lesser stress than in the case of a
surge with a vertical front.
The differential equation governing the variation of the voltage would be
a fourth order partial differential equation.
F
i
g
u
r
e

6
.
-
v
a
ri
a
ti
o
n
s
o
f
v
o
lt
a
g
e
a
n
d
v
o
lt
a
g
e
g
r
a
d
i
e
n
t
414 J R Lucas
High Voltage Testing
6.5Tests on Insulators
The tests on insulators can be divided into three groups.
These are the type tests, sample tests and the routing tests.
6.5.1 Type Tests
These tests are done to determine whether the particular design is suitable
for the purpose.
(a) Withstand Test:
The insulator should be mounted so as to simulate practical conditions.
A 1/50 μs wave of the specified voltage (corrected for humidity, air
density etc.,) is applied.
Flashover or puncture should not occur.
[If puncture occurs, the insulator is permanently damaged].
The test is repeated five times for each polarity.
(b) Flash-over test:
F
i
g
u
r
e

6
.
-
v
a
ri
a
ti
o
n
s
o
f
v
o
lt
a
g
e
a
n
d
v
o
lt
a
g
e
g
r
a
d
i
e
n
t
415 J R Lucas
High Voltage Testing
A 1/50 μs wave is applied.
The voltage is gradually increased to the 50% impulse flash-over voltage.
The test is done for both polarities.
There should be no puncture of insulation during these tests.
(c) Dry One-minute test:
The insulator, clean and dry, shall be mounted as specified and the
prescribed voltage (corrected for ambient conditions) should be gradually
brought up (at power frequency) and maintained for one minute.
There shall not be puncture or flash-over during the test.
Dry flash-over test: The voltage shall then be increased gradually until
flash-over occurs. This is repeated ten times. There shall be no damage
to the insulator.
(d) One-minute Rain test:
416 J R Lucas
High Voltage Testing
The insulator is sprayed throughout the test with artificial rain drawn from
a source of supply at a temperature within 10
o
C of the ambient
temperature of the neighbourhood of the insulator.
The rain is sprayed at an angle of 45
o
on the insulator at the prescribed
rate of 3 mm/minute.
The resistivity of the water should be 100 ohm-m ± 10%.
The prescribed voltage is maintained for one minute.
Wet flash-over test:
The voltage shall then be increased gradually until flash-over occurs.
This is repeated ten times.
There shall be no damage to the insulator.
(e) Visible discharge test:
417 J R Lucas
High Voltage Testing
This states that after the room has been darkened and the specified test
voltage applied, after five minutes, there should be no visible signs of
corona.
6.5.2 Sample Tests
The sample is tested fully, up to and including the point of breakdown.
This is done only on a few samples of the insulator.
(a) Temperature cycle test:
The complete test shall consist of five transfers (hot-cold-hot-....), each
transfer not exceeding 30 s.
(b) Mechanical loading test:
The insulator shall be mechanically loaded up to the point of failure.
When failure occurs, the load should not be less than 2000 lbf.
(c) Electro-mechanical test:
418 J R Lucas
High Voltage Testing
The insulator is simultaneously subjected to electrical and mechanical
stress. (i.e. it shall be subjected to a power frequency voltage and a tensile
force simultaneously.
The voltage shall be 75% of dry flash-over voltage of the unit. There
should be no damage caused.
(d) Overvoltage test:
The insulator shall be completely immersed in an insulating medium (oil),
to prevent external flashover occurring.
The specified overvoltage must be reached without puncture.
The voltage is then gradually increased until puncture occurs.
(e) Porosity test:
419 J R Lucas
High Voltage Testing
Freshly broken pieces of porcelain shall show no dye penetration after
having been immersed for 24 hours in an alcoholic mixture of fushing at a
pressure of 2000 p.s.i.
6.5.3 Routine Tests
These are to be applied to all insulators and shall be commenced at a low
voltage and shall be increased rapidly until flash-over occurs every few
seconds.
The voltage shall be maintained at this value for a minimum of five
minutes, or if failures occur, for five minutes after the last punctured piece
has been removed.
At the conclusion of the test the voltage shall be reduced to about one-
third of the test voltage before switching off.
420 J R Lucas
High Voltage Testing
Mechanical Routine Test: A mechanical load of 20% in excess of the
maximum working load of the insulator is applied after suspending the
insulator for one minute.
There should be no mechanical failure of the insulator.
421 J R Lucas
High Voltage Testing
6.6Tests on Transformers
The following sequence of tests is generally adopted for transformers.
System Voltage
I.E.C.
Impulse Withstand Voltage
11 kV 75 kV
33 kV 170 kV
66 kV 325 kV
132 kV 550 kV
275 kV 1050 kV
(1) Apply full wave impulse at 75% I.E.C. withstand value. Since the
transformer should be able to withstand the I.E.C.voltage, there should
be no damage to the transformer.
422 J R Lucas
High Voltage Testing
The values of R and C in the impulse generator are adjusted after
deriving to get the required waveform.
(2) Apply full wave at 100% I.E.C. withstand value and observe whether
there is any breakdown.
The waveform observed should be identical to applied waveform (other
than for amplitude) : then the device has passed the test.
(3) Chopped wave test at 115% fullwave amplitude : For this kind of test ,
the impulse generator would have to be fitted with a rod gap or controlled
trigatron type gap.
Since there is no voltage across insulator after chopping takes place, from
the waveform it is not possible to say whether any damage has taken
place.
(4) Therefore apply full wave test again and compare the wave and at 100%
of I.E.C. voltage and see whether there is any distortion in the waveform
indicating damage.(same as test 2)
423 J R Lucas
vvvvvttttt75
%
100
%
115
%
100
%
100
%
due to partial
discharge
(1
)
(2
)
(3
)
(4
)
(4)
*
High Voltage Testing

Figure 6.14 – Test Waveforms for transformers
424 J R Lucas
High Voltage Testing
425 J R Lucas
High Voltage Testing

F
i
g
u
r
e

6
.
-
c
h
o
p
p
e
d
w
a
v
e
f
o
r
m

a
t
1
5
0
%

v
o
lt
a
g
e
100% 150% 0t
426 J R Lucas
High Voltage Testing
Since the chopped wave test exerts considerable stress on the winding,
there is some controversy on the requirement of this test.
Thus the chopped wave requirement is not universal.
In the American industry, the chopped wave is conducted at 150% full
wave and such that the chopping is done at less than the peak value.
In this case the stress might in fact be very much more than in the British
method.
F
i
g
u
r
e

6
.
-
c
h
o
p
p
e
d
w
a
v
e
f
o
r
m

a
t
1
5
0
%

v
o
lt
a
g
e
427 J R Lucas
High Voltage Testing
6.7Tests on Cables
For cables not in the supper voltage class the tests to be carried out are
laid down in the appropriate British standard specifications.
Thus for paper impregnated insulated cables with lead or alloy sheaths,
BS 480 Part I: 1954, the tests (purely electrical) are as follows;
(1) Acceptance Tests at Works
(a) Conductor resistance
(b) Voltage test:
The applied voltage must be of approximately sinusoidal shape and of any
frequency between 25 and 100 Hz.
It must be increased gradually to the full value and maintain continuously
for 15 minutes between conductors and between each conductors and
sheath.
F
i
g
u
r
e

6
.
-
c
h
o
p
p
e
d
w
a
v
e
f
o
r
m

a
t
1
5
0
%

v
o
lt
a
g
e
428 J R Lucas
High Voltage Testing
The required values of the test voltages are tabulated in the specification
and, as one illustration of the magnitude relative to the normal voltages,
the figures for the 11 kV cables for earthed system are given in the table.
Voltage
Belted Cables
Single-core, S.L.
& Screened Cables
Designation (i) (ii) (ii)
(1) (2) (1) (2) (1) (2)
11 kV 24 kV 36 kV 14 kV 21 kV 15 kV 22 kV
where (i) Between conductors,
(ii) Between any conductor and sheath
(1) Cable as manufactured,
(2) After bending test
It will be seen that a voltage test is made before and after a bending test.
429 J R Lucas
High Voltage Testing
In this the cable has to be bent around a cylinder of specified diameter to
make one complete turn: it is then unwound and the process repeated in
the opposite direction .
The cycle of process has to be carried out three times.
(c) Dielectric power–factor / Voltage test (for 33 kV cables only) :
Each core of every drum of completed cable is tested for dielectric power
factor at room temperature at the following a.c. single phase 50 Hz
voltages : 9.5 kV , 19 kV , 28.5 kV, 38.0 kV.
The measured power–factor at normal working voltage shall not the value
declared by the manufacturer and shall in no case exceed 0.01 .
Ionization – ie. difference in power–factor between half the normal working
voltage and twice the normal working voltage – shall not exceed the value
declared by the manufacture and shall no case exceed 0.0006 for 3–core
screened cable or 0.001 for single core and screened S.L.type cable.
430 J R Lucas
High Voltage Testing
The manufacturer can also be asked to produced evidence to show that the
power factor at normal working voltage does not exceed 0.01 at series of
temperature ranging from 15
o
C to 65
o
C.
(2) Sample test at works
These include bending test above and a dripping or drainage test for
cables which have to be installed vertically.
(3) Test when installed
A voltage test similar to the above is carried out in the same manner but
with some what reduced voltages.
Thus the value of 24 kV, 14 kV and 11 kV for belted cable as
manufactured and the value 15 kV for single core, S.L. and screened
cables, become 20 kV, 11.5 kV and 12 kV respectively.
6.7.1 Tests on Pressurised Cables
431 J R Lucas
High Voltage Testing
Type approval tests, are stipulated for each design of cable and accessory.
These tests are carried out on the maximum and the minimum conductor
sizes for each design and voltage rating, and if successful, no further type
tests are required, except in the case of changes in the design.
The dielectric thermal resistance test included in the schedule is applied
only to the minimum conductor sizes.
The tests are as follows:
(a) Loading cycle test:
A test loop, comprising the cable and each type of accessory to be
subjected to 20 load cycles to a minimum conductor temperature 5
o
C in
excess of the design value, with the cable energised to 1.5 times the
working voltage.
The cable to be tested at a stipulated minimum internal pressure.
(b) Thermal stability test (132 kV cables only):
432 J R Lucas
High Voltage Testing
After test (a), the cable to be energised to 1.5 times working voltage and
the loading current adjusted to give a maximum temperature 5
o
C in
excess of the design value.
The current to be maintained at this value for a period of 6 hours, with
other test conditions unaltered, to prove that the cable is thermally stable.
For 275 kV cables, 1.33 times the working voltage is proposed.
(c) Impulse test:
A test loop, comprising cable and each type of accessory to be subjected
to 10 positive and 10 negative impulses at test voltage.
[Ex: Working voltage 132 kV, Impulse test voltage 640 kV, Peak working
voltage ratio during impulse test 6.0]
(d) Cold power-factor/voltage test:
433 J R Lucas
High Voltage Testing
The power factor of a 100 m length of cable to be measured at 0.5, 1.0,
1.5 and 2 times the working voltage with the cable at the stipulated
minimum internal pressure.
The values not to exceed the makers' guaranteed values.
(e) Dielectric thermal resistance test: The thermal resistance of the
cable is measured.
(f) Mechanical Test of metallic reinforcement: A sample of cable to
withstand twice the maximum specified internal pressure for a period of
seven days.
(g) Binding test: The cable to be subjected to three binding cycles round
a drum of diameter 20 times the diameter of the pressure retaining sheath.
The sample then to withstand the routine voltage test carried out on all
production lengths of cable.
6.8Tests on High Voltage Bushings
434 J R Lucas
High Voltage Testing
6.8.1 Bushing
A single or composite structure carrying a conductor or providing passage for
a conductor, through a partition, such as a wall or tank cover, or through a
ring type current transformer and insulating it there from, it includes the
means of attachment to the partition.
(i) Solid Bushing: A bushing consisting of a single piece of solid
insulating material which is continuous between its outer surface and the
inner conducting surface, which may be the main conductor or a conducting
layer connected thereto.
(ii) Plain Bushing: A bushing consisting of a single piece of solid
insulating material, with a space between the conductor and the inner surface
of the solid insulation.
The space is occupied by air, oil or other insulating medium which forms
part of the insulation. [See item (iii)]
(iii) Oil filled Bushing: A bushing consisting of an oil-filled insulating
shell, the oil providing the major radial insulation.
435 J R Lucas
High Voltage Testing
[Note: The conductor may be further insulated by a series of spaced
concentric cylinders which may be provided with cylindrical conducting
layers with the object of controlling the internal and external electric
fields.]
(iv) Condenser busing: A bushing in which cylindrical conducting layers
are arranged coaxially with the conductor within a solid body of insulating
materials, (including materials impregnated with oils or other
impregnants),the lengths and diameters of the cylinders being designed
with the object of controlling the internal and external electric fields.
[Note:A conductor bushing may be provided with a weather shield, in
which case the intervening space may be filled with oil or other insulating
medium.it is recommended that the term condenser bushing with oil
filling be used for this type.]
6.8.2 Tests on Bushings
436 J R Lucas
High Voltage Testing
Rating of bushings: Some of the relevant clauses from the standard is
given in the following sections.
Clause 4: A bushing shall be rated in terms of the following:
a) voltage(refer table 1, clause 5)
b) normal current (refer tables 2 and 3 clause 6)
c) frequency (refer clause 7)
d) insulation level (see clause 8 below)
Clause 8: The insulation level of bushing is designed by a voltage which
the bushing must be capable of withstanding under the specified test
conditions
For impulse tested bushings the rated insulation level is expressed as an
impulse voltage value i.e. impulse withstand voltage with 1/50 μs full wave.
For non–impulse tested bushings the rated insulation level is expressed as
a power frequency voltage value i.e. one minute dry withstand voltage.
437 J R Lucas
High Voltage Testing
Type Tests
Clause 14: Power frequency test
Clause 15: Impulse test
Clause 17: Momentary dry withstand test (power frequency voltage)
Clause 18: Visible discharge test (power frequency voltage)
Clause 19: Wet withstand test (power frequency voltage)
Clause 20: Puncture withstand test (power frequency voltage)
Clause 21: Full wave withstand test (impulse voltage)
Clause 22: Puncture withstand test (impulse voltage)
Sample Tests
Clause 23: Temperature rise test
438 J R Lucas
High Voltage Testing
Clause 25: Thermal stability test
Clause 26: Temperature cycle test
Clause 27: Porosity test
Routine Tests
Clause 29: One minute dry withstand test (power frequency voltage)
Clause 30: Oil lightness test
Clause 31: Power factor voltage test
6.9 Tests on Porcelain and toughened glass insulators for
overhead power lines
439 J R Lucas
High Voltage Testing
Specifications B.S.137:1960 (3.3 kV and upwards)
Classification of tests
Tests are divided into three groups, as shown.
Tests in Group I (Type tests)
These tests are intended to verify those characteristics of an insulator or
set, pin or line post insulator which depend on shape and size of the
insulator and of its metal parts and accessories.
They are normally made once only to establish design characteristics.
Clause 18: Impulse withstand voltage tests and 50% dry impulse
flashover test
Clause 19: Power frequency voltage one-minute wet test and wet
flashover test
Clause 20: Visible discharge test
440 J R Lucas
High Voltage Testing
Tests in Group II (Sample tests)
These tests are for the purpose of verifying certain characteristics of a
string insulator unit,line post insulator or pin insulator and pin and the
quality of the materials used.
They are made on insulators taken at random from every batch offered for
acceptance.
Clause 23: Verification of dimensions
Clause 24: Temperature cycle test
Clause 25: Mechanical failing load test or
Clause 26: Electro-mechanical failing load test
Clause 27: Overvoltage tests
Clause 28: Porosity test on porcelain insulators
441 J R Lucas
High Voltage Testing
Clause 29: Thermal shock test on toughened glass insulators: The glass
shall not shatter when the sample insulators are completely
immerses in water at a temperature not exceeding 50 C, the
temperature of the insulators immediately before immersion being
at least 100 C higher than that of the water.
Clause 30: Galvanizing test: The galvanized samples shall be tested in
accordance with B.S.729 and shall satisfy the requirements of that
standard.
Tests in Group III (Sample tests)
These tests are for the purpose of eliminating insulators with
manufacturing defects.
They are made on every insulator offered for acceptance.
442 J R Lucas
High Voltage Testing
Clause 32: Electrical test on porcelain insulators
Clause 33: Thermal shock test on toughened glass insulators
Clause 34: Mechanical test on string insulator units: Every string unit
shall be subjected to a tensile load of at least 40% of the specified
minimum failing load, for a period of not less than 10 seconds.

1.

High Voltage Breakdown Phenomena [6 hrs] Breakdown Characteristic in gases: Electron Avalanche Mechanism, Townsend Breakdown Process, Streamer Mechanism, Time lags of Spark breakdown. Corona Discharges, Mechanism of corona formation, Power Loss due to Corona Breakdown in Liquids: Breakdown of Commercial liquids; Breakdown due to gaseous inclusions, liquid globules, solid particles; Breakdown of Solid Insulating Materials: Electro-mechanical breakdown, Breakdown due to internal discharges, Surface Breakdown, Thermal Breakdown, Electrochemical Breakdown, Chemical Deterioration, Breakdown of Composite Insulation.

.

High Voltage Cables [6 hrs]
Power loss in the cable: Dielectric loss, Conductor loss, Sheath loss, Intersheath Loss, Cross-bonding of Cables Impregnated paper insulation: Properties required, Principle underlying the design, Paper insulated power cables, Single core and three core cables: Insulation Resistance, Capacitance, Copper Space Factor Dielectric stress in a single core cable: Cable Grading for Uniform Stress Distribution, Capacitance Grading, Intersheath Grading Pressurised high voltage cables: Oil-pressure cables, Gas-pressure cables, External Pressure Cables, Internal Pressure Cables Thermal design of cables: Current rating, Thermal Resistance single-core cables, three-core cables, protective coverings, ground around cable, Cables exposed to air High voltage bushings: Simple cylindrical bushing, Condenser bushing

3.

High Voltage Generators for Testing [6 hrs]
Generation of High Alternating Voltages:

Cascade arrangement of transformers Resonant Transformers High frequency high voltages
Generation of High Direct Voltages:

Rectifier circuits Voltage Multiplier Circuits Electrostatic generators: Van de Graeff generator, Sames Generator

4.

High Voltage Measurements and Testing [6 hrs] Electrostatic voltmeter, sphere gaps, potential dividers, matching, peak reading meters, Klydonograph Type tests, Sample Tests, Routine Tests Oscilloscopes for the measurement of fast transients Measurements of capacitance and loss tangent: High Voltage Schering Bridge, Dielectric loss measurement, Detection of internal discharges Measurement of dielectric constant and dissipation factor of a liquid dielectric General tests carried out on High voltage equipment. Testing of solid dielectric materials.

1
Breakdown of Gaseous Insulation

3

Electrical Insulating Materials referred to as Dielectrics electrostatic fields can remain almost indefinitely offer very high resistance to the passage of direct currents cannot withstand an infinitely high voltage.

J R Lucas

Breakdown of Gaseous Insulation

when applied voltage across dielectric exceeds a critical value the insulation will be damaged

may be gaseous, liquid or solid in form.

1

J R Lucas

Breakdown of Gaseous Insulation

Can a person swim one length without hitting some one ? If the pool is almost empty If the pool is crowded Will he be able get sufficient momentum in these cases to hurt someone?

Ionisation of Gases Gaseous dielectrics are not free of electrically charged particles, including free electrons. Free electrons may be caused by irradiation or field emission can lead to a breakdown process to be initiated. on the application of an electric field are accelerated from cathode to anode by the electric stress applying a force on them. Force = mass× acceleration, Force = charge× electric field acquire a kinetic energy (½ mu2) as they move through the field. moving towards the anode, collide with gas molecules present between the electrodes. part of the kinetic energy of the electrons is lost in these collisions, and part is transmitted to the neutral molecule.
1.1

1

J R Lucas

Breakdown of Gaseous Insulation

Energy is usually expressed as a voltage, Ei = e Vi in electron-volt (eV) as the energies involved are extremely small, where e is the charge on an electron = 1.6 x 10-19 C. 1 e V = 1.6 x 10-19 J If molecule gains sufficient energy (more than the ionisation energy Ei), it may ionise by collision. Mean number of ionising collisions by one electron per unit drift across the gap is not a constant but subject to statistical fluctuations. The newly liberated electron and the impinging electron are then accelerated in the field and an electron avalanche is set up.

3

J R Lucas

Breakdown of Gaseous Insulation

at voltages well below breakdown.3 J R Lucas Breakdown of Gaseous Insulation Further increase in voltage results in additional ionising processes. until ultimately breakdown occurs. constituting the well known corona discharge. Ionisation increases rapidly with voltage once these secondary processes take place. . however. In non-uniform fields. considerable ionisation may be present in the region of high stress. It is worth noting that in uniform fields. the ionisation present at voltages below breakdown is normally too small to affect engineering applications.

ions and photons with gas molecules. Since an electron in the outermost orbit is subject to the least attractive force from the nucleus.1.1. it is the easiest removed by any of the collision processes.1 Ionisation processes in gas discharges Electrical breakdown of a gas is caused by various processes of ionisation. and electrode processes which take place at or near the electrode surface [Electrons can be emitted from the cathode at stresses around 100 – 1000 kV/cm due to field emission]. Ionisation is the process by which an electron is removed from an atom. leaving the atom with a nett positive charge (positive ion). gas processes involving the collision of electrons. 1 J R Lucas Breakdown of Gaseous Insulation .

The energy required to remove an outer electron completely from its normal state in the atom to a distance well beyond the nucleus is called the first ionisation potential. In this case. a photon will be emitted having the same energy as previously absorbed. The reciprocal process of an electron falling from a great distance to the lowest unoccupied orbit is also possible. 1 J R Lucas Breakdown of Gaseous Insulation .

(i.1.1. in collision with a neutral gas molecule exceeds the ionisation energy (Ei = e Vi) of the molecule. When the gas molecules are bombarded with electrons. The probability of this process is zero for electron energies equal to the ionisation energy Ei.(½ mu²)  + + 2 eM In general.e. but increases almost linearly at first. the necessary but not necessarily sufficient condition is ½ mu² > Ei) M + e. and then gradually with electron energy up to a maximum. 3 J R Lucas Breakdown of Gaseous Insulation .2 Relevant gas ionisation processes (i) Ionisation by simple collision When the kinetic energy of an electron (½ mu²). other electrons bound to atoms may be freed by the collision with the high energy electron. a positive ion and 2 slow moving electrons will result. then ionisation can occur.

1 .i e 6 5 4 3 2 1 0 Figure 1.Ionisation probability curve in air o l 0 n e 0 i c s t a r t o i n o n e n p e r o g b y a b ( i e l V i ) t y 3 J R Lucas Breakdown of Gaseous Insulation .

and thus the probability of ionisation is small. The molecule may be left in an excited state M*. For much higher values of primary energies. This is maximum at primary electron energies of about 200 – 500 eV. the energy of the impinging electron would be sufficient for this electron to penetrate the surface deeper into the molecule. the energy transferred may not be sufficient to cause electrons to escape from the surface of the molecules. depend mainly on the energy of the primaries. so that again the chance of escape of other electrons decreases. For lower energy values.Ratio of. the neutral gas molecule does not always gets ionised on electron impact. 4 J R Lucas Breakdown of Gaseous Insulation . the electrons given by collision to the primary electrons. with energy Ee. (ii) Excitation In the case of simple collision.

Energy is given out when the electron jumps from one orbit to the next. then ionisation of can occur by a collision with a relatively slow electron. .2 J R Lucas Breakdown of Gaseous Insulation M + e.(½ mu²) → M* + eThis excited molecule can subsequently give out a photon of frequency v with energy emitted h v. M*  M + h v where h = Planck's constant = 6.624 x 10-34 J s (iii) Ionisation by Double electron impact If a gas molecule is already raised to an excited state (energy Ee) by a previous collision.

then a colliding electron may 3 J R Lucas Breakdown of Gaseous Insulation .e. where h = Plank's constant = 6. h ν > Ei . (i.624 x 10-34 joule) M + h  → M+ + e(v) Electron Attachment If a gas molecule has unoccupied energy levels in its outermost group.Less energy would be required than the ionisation energy.Ee) M* + e.e.(½ mu²) → M+ + 2 e(iv) Photo-ionisation A molecule in the ground state can be ionised by a photon of frequency  provided that the quantum of energy emitted h ν (by an electron jumping from one orbit to another). necessary but not sufficient condition is ½ mu² > Ei . (i. is greater than ionisation energy of molecule. but the energy must exceed the additional energy required to attain the ionisation energy.

and becomes a neutral molecule. caused by the excess energy. converting the molecule into a negative ion M-. M. M + e. Note: Electron attachment decreases the number of free electrons.→ MThe negative ion thus formed would be in an excited state. 2 J R Lucas Breakdown of Gaseous Insulation .→ M + e(vii) Other Processes The above discussed processes are the most important in relation to the gas discharge phenomena. unlike ionisation which increases the free electrons. (vi) Electron detachment This occurs when a negative ion gives up its extra electron.take up one of these levels.

and atom-atom collisions.Other possible gas processes include ion-atom collisions. 1 J R Lucas Breakdown of Gaseous Insulation . Normally ions and atoms having such energies are encountered only in high current arcs and thermonuclear discharges. a positive ion must possess energy of at least 2 eV. excited atommolecule collisions. which allows the internal motion of the atomic system to adjust itself gradually to the changing condition without any energy transition occurring. In order to cause ionisation of a neutral unexcited atom of its own kind. due to the relatively slow interaction time. It should be noted that collisions between ions and atoms rarely result in ionisation.

2.1 Electron Avalanche Mechanism Suppose a free electron exists (caused by some external effect such as radio–activity or cosmic radiation) in a gas where an electric field exists. Both depend on an initial electron avalanche mechanism.1. 3 J R Lucas Breakdown of Gaseous Insulation . then it is likely to ionize a gas molecule by simple collision resulting in 2 free electrons and a positive ion.2 Breakdown Characteristic in gases Two mechanisms of breakdown in gasses are known which explain the behaviour under different conditions. These are the Townsend and streamer mechanisms. These 2 electrons will be able to cause further ionization by collision leading in general to 4 electrons and 3 positive ions. 1. If the field strength is sufficiently high.

3 J R Lucas Breakdown of Gaseous Insulation Molecule – A Cathode Anode C ½mu2 + F Electron n a i Positive ion o t d h g e o u d r + e e – 1 . A swarm of electrons and positive ions produced in this way is called an The v electron avalanche. and the number of free electrons will go on – increasing as they continue to move under the action of the electric field. a l a n . 2 The process is cumulative.

it may grow until it contains many P millions of electrons. 1 J R Lucas Breakdown of Gaseous Insulation .c h e r o c e s s b u i l d u p In the space of a few millimetres.

Let n0 = No.2. on average. made by one electron per unit drift in the direction of the field. of electrons/second moving at a distance x from cathode [nx > n0 due to ionising collisions in gap] α = number of ionising collisions.1.1 Mathematical Analysis [Townsend Breakdown Process ] When the voltage applied across a pair of electrodes is increased. Impact ionization by electrons is probably the most important process in the breakdown of gasses. nx = No. of electrons/second emitted from the cathode. but this process alone is not. the nonself-sustaining current throughout the gap increases slowly. The electrons emitted from the cathode move through the gas with an average velocity determined by their mobility for the field strength existing for the particular value of voltage. [Townsend's first ionisation coefficient] 3 J R Lucas Breakdown of Gaseous Insulation .1.

The nx electrons entering the laminar will traverse it in the presence of the applied field E. Consider a laminar of thickness dx at a distance x from the cathode. .C Anode + d x a x t h o d e – 3 J R Lucas Breakdown of Gaseous Insulation then 1/α= average distance traversed in the field direction between ionising collisions. The ionising collisions generated in the gas gap will be proportional to both dx and to nx.

Thus Therefore 5 dnx ∝ nx ∝ dx J R Lucas Breakdown of Gaseous Insulation dnx = α . dx (from definition of α) . nx .

on the average. . x n0 ∫ nx d nx = α ∫ dx x = n0 . then the number of electrons nd striking the anode per second is given by nd = n0 . the number of positive ions arriving at the cathode/second must be exactly equal to the number of newly formed electrons arriving at the anode. eαd Therefore.n0)/n0 new electrons (and corresponding positive ions) in the gap. In the steady state. eα x nx If the anode is at a distance x = d from the cathode. each electron leaving the cathode produces (nd .Rearranging and integrating gives 1 J R Lucas Breakdown of Gaseous Insulation nx 0 loge ( n x / n0 ) = α .

the electron impact ionization is attended by secondary processes on the cathode. Consider the current growth equations with secondary mechanism also present. [Townsend's second ionisation coefficient] n0 = No. of primary photo-electrons/second emitted from the cathode n0' = No. with every newly formed avalanche surpassing the preceding one in the number of electrons. of secondary electrons (on average) produced at the cathode per ionising collision in the gap. In the actual breakdown process. of secondary electrons/second produced at the cathode 1 J R Lucas Breakdown of Gaseous Insulation . eαd where I0 is the initial photo-electric current at the cathode. which replenish the gas gap with free electrons. Let γ = No.Thus the circuit current will be given by I = I0 .

1 ) n 0" = n0 + n 0' αd n 0" = n 0 + n 0" ( e .3 n0" = Total number of electrons/second leaving the cathode n0" = n0 + n0' J R Lucas Breakdown of Gaseous Insulation Then On the average.1).1 ) .1] collisions in the gap. giving the number of ionising collisions/second in the gap as n0" (eαd . Thus by definition n0′ ' γ= αd n0" ( e . each electron leaving the cathode produces [eαd . γ .1 ) giving but so that αd n0′ = γ n 0" ( e .

1 ) αd Similar to the case of the primary process (with α only). .1 J R Lucas Breakdown of Gaseous Insulation This gives the result n 0" = n0 1 . when the denominator of the circuit current expression becomes zero and the current I → ∞ . we have n0 e αd nd = n0" e = 1 . in steady state. the circuit current I will be given by I0 e I= 1 . eαd and γ eαd increase until γ eαd → 1 .γ (eα d .γ ( eα d .1 ) αd This equation describes the growth of average current in the gap before spark breakdown occurs.1 ) Thus.γ ( eα d . As the applied voltage increases.

In this case. typically over 1 μs and does not generally occur with impulse voltages. in practice.1) = 1 This condition is known as the Townsend criteria for spark breakdown. This condition may thus be defined as the breakdown and can be written as γ (eαd . the current will. 1 J R Lucas Breakdown of Gaseous Insulation . be limited only by the resistance of the power supply and the conducting gas. The avalanche breakdown develops over relatively long periods of time.

The voltage current characteristics are then obtained for different gap settings. and the electrode system consists of a plane high voltage electrode and a low voltage electrode surrounded by a guard electrode to maintain a uniform field.2 Determination of Townsend's Coefficients α and γ Townsend's coefficients are determined in an ionisation chamber which is first evacuated to a very high vacuum of the order of 10-4 and 10-6 torr before filling with the desired gas at a pressure of a few torr. 3 J R Lucas Breakdown of Gaseous Insulation .1. At low voltage the current growth is not steady.2. The cathode is irradiated using an ultra-violet lamp from the outside to produce the initiation electron.1. The applied direct voltage is about 2 to 10 kV. The low voltage electrode is earthed through an electrometer amplifier capable of measuring currents in the range 0.01 pA to 10 nA.

1 J R Lucas Breakdown of Gaseous Insulation .3.Afterwards the steady Townsend process develops as shown in figure 1.

1 J R Lucas Breakdown of Gaseous Insulation Disch non-self selfV 1.3 sustaining Figure brea o arge Current V sustaining discharges Typical kdo 2 curre discharges growth wn nt I V1 voltage V .

the discharge current is given by I= I0 e ≈ I 0 eα d when α d » 1 1 .From the Townsend mechanism.1 ) αd 1 J R Lucas Breakdown of Gaseous Insulation This can be written in logarithmic form as ln I = α d + ln I0 y = mx + c .γ ( eα d .

4 -αIo gradient =ln Variation of ln I d d vs .1 J R Lucas Breakdown of Gaseous Insulation Intercept = ln I gap spacing Figure 1 α d <<1.

3 J R Lucas Breakdown of Gaseous Insulation .From a graph of ln I vs d.2.1. The gases in which electron attachment occurs are electro-negative gases. the constants α and I0 can be determined from the gradient and the intercept respectively. γ can be determined from points on the upward region of the curve. The experiment can be repeated at different pressures if required. electron attachment to neutral molecules was not considered. 1. as in figure 1. Electron attachment removes free electrons and thus gives gases very high dielectric strengths. Once I0 and α are known.3 Breakdown in electronegative gases In the above analysis.4.

1] 1. Under these conditions.η e α -η   I = I0  α [e( α -η ) d .  α ( α -η ) d η  α . as the number of attachments per electron per unit drift in the direction of the field.An attachment coefficient η can be defined.γ α -η 3 J R Lucas Breakdown of Gaseous Insulation The corresponding criteria for spark breakdown is α γ [ e( α -η ) d − 1] = 1 α -η . the equation for the average current growth in a uniform field can be shown to be as follows. analogous with α.

1. Thus applying Townsend's criterion for spark breakdown of gases gives γ ( eα d .2. their mean energies attain equilibrium values dependant on the ratio E/p.1 ) = 1 which may be written in terms of the functions as  V  p d f 1     e f 2   p d    V pd     − 1  =1   . or more precisely α/p = f1(E/p) and γ = f2(E/p) For a uniform field gap.2 Paschen's Law 1 J R Lucas Breakdown of Gaseous Insulation When electrons and ions move through a gas in a uniform field E and gas pressure p. the electric field E = V/d.

d) In the above derivation the effect of temperature on the breakdown voltage is not taken into account.e. This is the statement of Paschen's Law. Thus it can be deduced that the correct statement of the above expression is V = f(ρ. where ρ is the gas density.d). i. . Using the gas equation pressure × volume = mass R × absolute Temperature pressure = density  R × absolute Temperature. V = f (p.3 J R Lucas Breakdown of Gaseous Insulation This equation shows that the breakdown voltage V is an implicit function of the product of gas pressure p and the electrode separation d.

 d where d is the gap spacing For air.5 shows a typical breakdown vs spacing characteristic. under normal conditions.5 . .4 kV/cm and B = 6. but reduces to about 25 kV/cm for large gaps of several meters] Figure 1.1 J R Lucas Breakdown of Gaseous Insulation breakd 1. A = 24. it is experimentally found that the breakdown voltage of a uniform field gap may be expressed in the form V = A.d + B.Breakdown gap Figure own spacing d characteristic voltage (cm) (kV) Under constant atmospheric conditions. [The breakdown voltage gradient is about 30 kV/cm in uniform fields for small gaps of the order of 1 cm.29 kV/cm1/2.

However.d Figure 1. breakdown spacing pressure x .T. ( ρ d ) 2 ρ0 ρ0 1 2 1 J R Lucas Breakdown of Gaseous Insulation = A ( δ d )+ B ( δ d ) where δ = relative density (or gas density correction factor).1 mm at N.This variation with spacing can be modified using Paschen's Law to include the variation with gas density as follows.P. with very low products of pressure × spacing. known as the Paschen's minimum.Breakdown characteristic for low p. a minimum breakdown voltage occurs. V= 1 A B .6 voltage (kV) (cm) values This equation is true for gap spacings of more than 0. . ρ d + 1/2 .

6). it follows that a lower electric stress suffices to impart to electrons the kinetic energy (½ m u2) required to ionize by collision.This can be explained in the following manner. decreases and an electron moving in the field consequently makes fewer collisions with gas molecules as it travels towards the anode. and let the pressure decrease from a point to the right of the minimum (Figure 1. The density therefore. even if the energy of the electron exceeds the ionisation energy. It is necessary now to take into account the fact that an electron does not necessarily ionize a molecule on collision with it. Since each collision results in a loss of energy. 1 J R Lucas Breakdown of Gaseous Insulation . Near the minimum of the characteristic. the density is low and there are relatively few collisions. Consider a gap of fixed spacing.

The electrons have a finite chance of ionizing.P.T. breakdown can occur only if the chance of ionising is increased. at lower voltage) at longer distances. Paschen's Law fails. which depends upon its energy. Also.e. breakdown to the left of the minimum occurs more readily (i. At very low pressures. Typically the voltage minimum is 300 V and occurs at a product or p. and this accounts for the increase in voltage to the left of the minimum. or at a gap of about 0. Paschen's Law is valid for temperatures below about 11000C.06 mm at N. 1 J R Lucas Breakdown of Gaseous Insulation . and at very high pressures (compared with atmospheric).d of 5 torr mm. If the density and hence the number of collisions is decreased. It is worth noting that if the density is fixed.

and hence the breakdown stress required. operation is well to the right of Paschen's minimum. Thus both high vacuum as well as high gas pressure can be used as insulating media. under very high pressure and under very low pressures (high vacuum) conditions. is very much higher than at atmospheric conditions. the breakdown voltage.A further increase in temperature ultimately results in the failure of Paschen's Law because of thermal ionisation. Gas Vmin (V) p. that for a constant gap spacing. [The vacuum breakdown region is that in which the breakdown voltage becomes independent of the gas pressure] Under normal conditions.d at Vmin(torr-cm) 1 J R Lucas Breakdown of Gaseous Insulation . It can be seen from the breakdown characteristic.

7 –– + ++ – – + –catho – + + Streamer anod − − – – –+ ++ – –+ – ++ Mechanism – – – ++ ++ –de – – +e + + + ++ –+ – – + – – – +++ .5 0.9 1.6 Table 1.2 J R Lucas Breakdown of Gaseous Insulation Air Argon Hydrogen Helium Carbon dioxide Nitrogen Nitrous oxide Oxygen Sulphur dioxide Hydrogen Sulphide 327 137 273 156 420 251 418 450 457 414 0. + – Figure 1.1 .7 0.67 0.0 0.15 4.51 0.33 0.Minimum spark breakdown voltages The table 1.567 0.1 gives the minimum sparking potential for various gases.

The Streamer theory predicts the development of a spark discharge directly from a single avalanche. The space charge produced in the avalanche causes sufficient distortion of the electric field that those free electrons move towards the avalanche head. in actual practice many discharges are found to be filamentary and irregular. While the Townsend mechanism predicts a very diffused form of discharge.2.++ – – – – ++ ++ – – ++ ++ +++ + ++ ++ 2 ++ + ++ + ++ + ++ + + + + + + + 1.8 + +++ + ++ + + ++ Streamer anode + + cathod ++ + +++ Mechanism e ++ + + + +++ ++ + + + ++ ++ + + + + + ++++ + +++ + ++ + ++ + + +++ . + +++++ – Figure 1.3 Streamer Mechanism J R Lucas Breakdown of Gaseous Insulation This type of breakdown mainly arises due to the added effect of the space– charge field of an avalanche and photo–electric ionization in the gas volume. and in so doing generate further avalanches in a process that rapidly becomes cumulative.

the space charge increases.8 shows the streamer breakdown process.+ + ++ ++ + + 2 J R Lucas Breakdown of Gaseous Insulation As the electrons advance rapidly. 1.4 Factors affecting the breakdown voltage a Vacuum gap . Just behind the head the field between the electrons and the positive ions is in the opposite direction to the applied field and hence the resultant field strength is less.The field will be enhanced in front of the head. Figure 1. The process is very fast and the positive space charge extends to the cathode very rapidly resulting in the formation of a streamer. causing a further enhancement of the field around the anode.2. Again between the tail and the cathode the field is enhanced. (Figure 1. the positive ions are left behind in a relatively slow–moving tail.7) Due to the enhanced field between the head and the anode.

However. This gives a constant breakdown strength. 2 J R Lucas Breakdown of Gaseous Insulation . all other parameters remaining constant. (i) Electrode Separation For vacuum gaps less than about 1 mm. The breakdown voltage of a high vacuum is the voltage which when increased by a small amount will cause the breakdown of the gap that was held at that voltage for an infinite time.Vacuum is ideally the best insulator. the breakdown voltage is approximately proportional to the length. this definition is not always practicable as the breakdown is affected by many factors. with breakdown strengths of the order of 104 kV/c.

For these small gaps. V = k. being of the order of 1 MV/cm. the breakdown stress is relatively high.9 breakdo gap spacingvoltage-(kV) 1 stress breakdown if wn3 (cm) Breakdown 0 caused by field 4 voltage (kV) (kV/cm) emission only characteristic 50 0 1 J R Lucas Breakdown of Gaseous Insulation for d < 1 mm 25 0 0 . Field emission of electrons probably plays an important part in the breakdown process.d breakdown Breakdown 2 actual 75 Figure 1.

it is the product of breakdown voltage and breakdown stress that remains constant. so that V = k2 d1/2 Or in a more general form. V . being about 10 kV/cm at a spacing of 10 cm. For an Uniform field gap. E = V/d. both regions can be expressed by the equation 3 J R Lucas Breakdown of Gaseous Insulation .9. the breakdown voltage does not increase at an equal rate and the apparent breakdown stress for longer gaps is much reduced. where the constant k1 depends on the material and surface conditions of the electrodes. [Apparent stress is defined as the voltage required to cause breakdown divided by the distance between the electrodes] Figure 1. E = k1 for d > 1 mm. Cranberg has shown theoretically that for longer gaps.For gaps greater than about 1 mm.

V = k dx 3 where x = 0. until a constant value is reached.5 for long gaps > 1 mm = 1 for gaps < 1mm J R Lucas Breakdown of Gaseous Insulation (ii) Electrode Effects Conditioning The breakdown voltage of a gap increases on successive flashovers. . The electrodes are then said to be conditioned. the electrodes must first be conditioned in such a way that reproducible results are obtained. When investigating the effect of various factors on breakdown. This increase in voltage is ascribed to the burning off by sparking of microscopic irregularities or impurities which may exist on the electrodes.

10 5 Breakdo 10 – Breakdown 15 Figure of 20 wn 25 characteristic breakdo Voltage 40 wn (kV) 30 20 10 0 .3 J R Lucas Breakdown of Gaseous Insulation × Number1.

and hence breakdown values that are normally obtained experimentally are those of conditioned electrodes.The effect of conditioning is shown. (iii) Material and Surface Finish Electrode Material Steel Stainless Steel Nickel Monel metal Aluminium Voltage across 1 mm gap (kV) 122 120 96 60 41 1 J R Lucas Breakdown of Gaseous Insulation . The breakdown voltage of conditioned electrodes or gaps is very much reproducible than otherwise. Unconditioned electrodes may have breakdown values a low as 50% of the breakdown voltage with conditioned electrodes.

Effect of electrode material on breakdown The electrode surfaces form the physical boundaries between which the breakdown finally takes place. (iv) Surface contamination Presence of contamination in the test cell reduces the breakdown voltage sometimes by as much as 50% of the clean electrode value.2 J R Lucas Breakdown of Gaseous Insulation Copper 37 Table 1. In general. (v) Area and configuration of electrodes . Thus it is not surprising to find that the breakdown strength of a given size of gap is strongly dependant on the material of the electrodes.2 . the greater the breakdown voltage. the smoother the surface finish.

For example. whereas electrodes of the same material of area 1000 cm2 gives a breakdown voltage across the same 1 mm gap as 25 kV. electrodes of 20 cm2 area gives the breakdown voltage across a 1 mm gap of 40 kV. the more convex electrodes have higher breakdown voltage than the more nearly plane electrodes even though at the same voltage they carried a higher electric field at the surface. Cooling the electrodes to liquid Nitrogen temperature increases the breakdown voltage. (vii) Frequency of applied voltage 2 J R Lucas Breakdown of Gaseous Insulation . and for nickel and iron electrodes. the strength remains unchanged for temperatures as high as 5000C. Thus breakdown voltage decreases slightly with increase in surface area. Up to 1 mm gap.Increasing the area of the electrodes makes it more difficult to maintain a given breakdown voltage. (vi) Temperature The variation of the breakdown voltage with temperature is very small.

However. it has been shown that for a small gap (2 mm) there is no dependence of the breakdown voltage on the frequency in the range 50 Hz to 50 kHz. and a higher alternating voltage than a direct voltage. Figure Pressure 1200 1.11 – Breakdown characteristic 10-3 Breakdown -5 10-6 10 10-4 (torr) (kV) Voltage 1000 800 600 400 200 0 1 J R Lucas Breakdown of Gaseous Insulation .It is known that a given gap stands a higher impulse voltage than an alternating voltage.

For example.11 was obtained. the characteristics shown in figure 1. so that the breakdown stress at this stage could in fact be improved by actually worsening the vacuum. However.(viii) Vacuum Pressure 2 J R Lucas Breakdown of Gaseous Insulation For small gaps. .6 mm diameter sphere opposite a plane cathode and a gap of 200 mm. The vacuum breakdown region is the region in which the breakdown voltage becomes independent of the nature of the pressure of the gap between the electrodes. increasing the degree of vacuum increases the breakdown voltage until below a certain pressure there is no change. using a 1. for large gaps (about 200 mm spacing) it is found that below a certain pressure the breakdown voltage starts decreasing again.

but rose with increasing pressure to a maximum at 5 x 10 .4 torr.4 torr. with further increase in pressure the voltage fell sharply. 2 J R Lucas Breakdown of Gaseous Insulation .The breakdown voltage was constant for pressure less than 5 x 10 .

a certain time elapses before actual breakdown occurs even though the applied voltage may be much more than sufficient to cause breakdown.2. 1 J R Lucas Breakdown of Gaseous Insulation .5 Time lags of Spark breakdown In high voltage engineering.1. On the application of a voltage. In considering the time lag observed between the application of a voltage sufficient to cause breakdown and the actual breakdown the two basic processes of concern are the appearance of avalanche initiating electrons and the temporal growth of current after the criterion for static breakdown is satisfied. breakdown under impulse fields is of great importance.

depending on the gap volume. there is usually no difficulty in finding an initiatory electron from natural sources (ex.12 .3 J R Lucas Breakdown of Gaseous Insulation → V t 1. However. natural sources may not be sufficient to provide an initiating .Time lag of Figure time Breakdo ← Impulse Collapse impulse of Voltage (µ s) wnbreakdown Voltage Voltage Vs 0 In the case of slowly varying fields. cosmic rays. detachment of gaseous ions etc). for impulses of short duration (around 1 μs).

The time ts which elapses between the application of a voltage greater than or equal to the static breakdown voltage (Vs) to the spark gap and the appearance of a suitably placed initiatory electron is called the statistical time lag of the gap. 2 J R Lucas Breakdown of Gaseous Insulation . Figure 1. and is shown in the diagram. the appearance being usually statistically distributed.electron while the voltage is applied. After such an electron appears.12 shows the time lag ‘t’ for an impulse voltage waveform. The sum tf + ts = t is the total time lag. and in the absence of any other source. The ratio V/Vs. which is greater than unity. the time tf required by the ionisation processes to generate a current of a magnitude which may be used to specify breakdown of the gap is known as the formative time lag. is called the impulse ratio. and clearly depends on ts + tf and the rate of growth of the applied voltage. breakdown will not occur.

e. If β = rate at which electrons are produced in the gap by external irradiation P1 = probability of an electron appearing in a region of the gap where it can lead to a spark P2 = probability that such an electron appearing in the gap will lead to a spark 1 J R Lucas Breakdown of Gaseous Insulation . the breakdown may occur in the wavefront or the wavetail of the impulse waveform.The breakdown can also occur after the peak of the voltage pulse. (i) Statistical Time lag ts The statistical time lag is the average time required for an electron to appear in the gap in order that breakdown may be initiated. Depending on the time lag. on the wavetail). (i.

The most favourable position is. If the gap is overvolted. then the probability of producing a current sufficient to cause breakdown rapidly increases. β increases and therefore decreases. the average time lag 3 J R Lucas Breakdown of Gaseous Insulation ts = 1 β P1 P 2 ts If the level of irradiation is increased. the probability of an electron appearing in a favourable position to produce breakdown. with clean cathodes of higher work function β will be smaller for a given level of illumination producing longer time lags.then. . The type of irradiation used will be an important factor controlling P1. Also. of course near the cathode.

As P2 → 1 with increasing overvoltage. ts → 1/βP1. 3 J R Lucas Breakdown of Gaseous Insulation .P2 therefore increases with overvoltage and may tend to unity when the overvoltage is about 10%.

and the time rate of development of ionisation will depend on the particular secondary process operative. Here again. . The value of the formative time lag will depend on the various secondary ionisation processes. it can be assumed that the initiatory electron is available which will eventually lead to breakdown.(ii) Formative time lag 1 J R Lucas Breakdown of Gaseous Insulation After the statistical time lag. an increase of the voltage above the static breakdown voltage will cause a decrease of the formative time lag tf. An uninterrupted series of avalanches is necessary to produce the requisite gap current (μA) which leads to breakdown. The additional time lag required for the breakdown process to form is the formative time lag.

24 (kV) .(iii) Time lag characteristic Figure 1.Single core conductor channel cable V2 core conductor 1 J R Lucas Breakdown of Gaseous Insulation .Time-lag characteristics breakdown characteristic 3.13characteristic Static Breakdown ← time lag .

3 J R Lucas Breakdown of Gaseous Insulation The time lag characteristic is the variation of the breakdown voltage with time of breakdown.13. The time lag characteristic based on the impulse waveform is shown in figure 1. and can be defined for a particular waveshape. Figure 1.14 timeARod – Typical Time lag Transform Breakdo → Rod ← gap (ii) er characteristics breakdown (i) Protected (µ s) wn Unprotecte VA Voltage d 0 .

In such a case. protection will be offered only in the region where the rod gap characteristic is lower than that of the transformer. Thus it is usual to have the rod gap characteristic slightly higher (gap ii) resulting in the intersection of the characteristics as shown. However.The time lag characteristic is important in designing insulation. then the breakdown voltage characteristic of rod gap must be less than that of the transformer at all times (gap i) to protect it from dangerous surge voltages. 1 J R Lucas Breakdown of Gaseous Insulation . even due to the smallest overvoltages which would in fact cause no harm to the system. If a rod gap is to provide secondary protection to a transformer. Thus it is likely that there would be frequent interruptions. as the sharpness of the two characteristics are different. This will ensure that the gap will always flashover before the protected apparatus. the gap setting will be low. with such a rod gap.

On the other hand.6 Corona Discharges In a uniform electric field. it is better and cheaper to replace 1 transformer a year due to this decision than have to double the cost of each of 100 such transformers in the system.2. 1 J R Lucas Breakdown of Gaseous Insulation .] 1. if the field is non–uniform.This crossing point is found from experience for a value of voltage which is highly unlikely to occur. an increase in voltage will first cause a localised discharge in the gas to appear at points with the highest electric field intensity. The other alternative is of course to increase the transformer characteristic which would increase the cost of the transformer a great deal. [This decision is something like saying. namely at sharp points or where the electrodes are curved or on transmission line conductors. a gradual increase in voltage across a gap produces a breakdown of the gap in the form of a spark without any preliminary discharges.

This voltage is known as the disruptive critical voltage (dcv). initially nothing will be seen or heard.2. This voltage is the visual corona inception voltage. This phenomena is always accompanied by a hissing noise. 1.This form of discharge is called a corona discharge and can be observed as a bluish luminance.1 Mechanism of Corona formation on a 2 conductor line When a gradually increasing voltage is applied across two conductors.6. and at a certain voltage a hissing noise is heard caused by formation of corona. A further increase in the voltage would cause a visible violet glow around the conductors. 1 J R Lucas Breakdown of Gaseous Insulation . and the air surrounding the corona region becomes converted to ozone. air surrounding the conductors get ionised. As voltage is increased. Corona is responsible for considerable power loss in transmission lines and also gives rise to radio interference.

The formation of corona causes the current waveform in the line. In the case of alternating voltages. while the negative conductor will be more patchy and often accompanied by streamers at rough points. both conductors appear to have a uniform glow. the glow observed will be uniform on the positive conductor. but when observed stroboscopically the effect is seen to be similar to the direct voltage case. the corona increases and finally spark over would occur between the two conductors. 1 J R Lucas Breakdown of Gaseous Insulation . If the conductors are placed quite close together.If the applied voltage is direct. corona formation may not occur before the spark over occurs. The condition for this will be considered later. If the voltage is further increased. and hence the voltage drop to be non-sinusoidal.

To prevent the formation of corona. the velocity of the electrons in the vicinity of the line increases. and the electrons acquire sufficient velocity to cause ionisation. Thus steel cored aluminium has the advantage over hard drawn copper conductors on account of the larger diameter.It also causes a loss of power. Corona formation may be reduced by increasing the effective radius. 1 J R Lucas Breakdown of Gaseous Insulation . the working voltage under fair weather conditions should be kept at least 10% less than the disruptive critical voltage. other conditions remaining the same. There is always some electrons present in the atmosphere due to cosmic radiation etc. When the line voltage is increased.

The advantage of corona in this instance is that it reduces transients by reducing the effective magnitude of the surge by partially dissipating its energy due to corona. These cause interference in the immediate vicinity of the line. The current flowing into a corona discharge contains high-frequency components. the disturbing field makes its appearance long before corona loss becomes appreciable. As the voltage is gradually increased.The effective conductor diameter can also be increased by the use of bundle conductors. The effect of corona on radio reception is a matter of some importance. 1 J R Lucas Breakdown of Gaseous Insulation . Corona acts as a safety valve for lightning surges. by causing a short circuit.

The interference fails to about a tenth at 50 m from the axis of the line. 1 J R Lucas Breakdown of Gaseous Insulation .The field has its maximum value under the line and attenuates rapidly with distance.

6.1.2.2 Waveform of Corona Current 1 J R Lucas Breakdown of Gaseous Insulation .

3 J R Lucas Breakdown of Gaseous Insulation Corona Applied Sine DC D t C volta 1.15 Waveform of Figure current voltage wave current V ge corona .

and there is no power loss in the line under no-load conditions. When the applied voltage is increased and corona is formed. This resultant waveform is shown in figure 1. 5 J R Lucas Breakdown of Gaseous Insulation .The shunt current in a line is almost purely capacitive under normal conditions.15. Thus the current waveform would consist of two components. and leads the applied voltage by 900. the air is rendered conducting. The corona current can be analysed and shown to possess a strong third harmonic component. The shunt current would no longer be leading the voltage by 900. and power loss occurs. The lossy component would be non-sinusoidal and would occur only when the disruptive critical voltage is exceeded in either polarity.

Thus when the stress has been raised to critical value immediately surrounding the conductor. and secondly. The effect is to increase the effective conductor diameter while the voltage remains constant.2. ionisation would commence only in this region and the air in this region would become conducting.1. and decreases rapidly as the distance from the conductor increases. This results in two effects. this would cause a reduction of the effective spacing between the conductors leading to an increase in stress.6. 1 J R Lucas Breakdown of Gaseous Insulation . Firstly. an increase in the effective sharpness of the conductor would reduce the stress outside this region.3 Mechanism of corona formation The stress surrounding the conductor is a maximum at the conductor surface itself.

In stormy weather. the number of ions present is generally much more than normal.Depending on which effect is stronger. If the stress is made to increase. the stress at increasing distance can either increase or decrease. c r effective xξ d Figure 1. The condition for stable corona can be analysed as follows. This reduced voltage is generally about 80% of the fair weather voltage. the breakdown strength of air can be taken as 30 kV/cm. Under ordinary conditions. Corona will of course be affected by the physical state of the atmosphere. and corona will then be formed at a much lower voltage than in fair weather.16 . further ionisation would occur and flashover is inevitable.Electric Stress in two conductor system o n c 1 J R Lucas Breakdown of Gaseous Insulation .

Breakdown of Gaseous Insulation Thus the conductor to neutral voltage. e ε0 2π x l n where q is the charge on each conductor in length l. which is half the difference would be equal to this value. . dx 2 π x ε0 p r e Since both charges (+q and -q) produce equal potential differences. and n o aseparated from the return conductor by a distance d is given by r 1 q ξ = . v eThus the potential V can be determined from V = ∫ ξ dx l d -r q o V = ∫ .u r radius c o 2 J R Lucas tThe electric stress ξ at a distance x from a conductor of radius x. the total potential difference between the two conductors is double this value.

Thus the conductor to neutral voltage is q d -r  V = . loge   2π ε0  r  1 J R Lucas Breakdown of Gaseous Insulation Therefore the electric stress at distance x is given by ξx = ξx = V x loge V d x loge r d -r r if d >> r [ξx and V can both be peak or both rms] qa db = – qa – c qb + .

2 = E 0. the maximum stress will occur when x is a minimum.rms d r loge r 2 J R Lucas Breakdown of Gaseous Insulation . Total charge is zero. Thus if E0. Similarly is qb and – qb. Since there is no electric stress within the conductor. ξmax = 30 kV/cm. For air. when V is the voltage to neutral and d is the equilateral spacing. that is at x = r. These give the solution to the three phase case with equilateral spacing as the same as the single phase solution. ξ rms = 21.rms is the rms value of the disruptive critical voltage to neutral. so that qc = – qa – qb qa and – qa are a pair similar to the two conductor case.For three phase lines. with equilateral spacing.2 kV/cm. so that ξrms = 30/2 = 21.

= 0. it is more liable to corona. Since the corona formation is affected by the mean free path. and = 0.90 for smooth polished conductors. 1 J R Lucas Breakdown of Gaseous Insulation .83 for 7 strand cables. a correction factor δ is introduced. This air density correction factor is given by the usual expression.0 0.87 to 0. Thus an irregularity factor m0 is introduced to account for this reduction. with p being the pressure expressed in torr and t being the temperature expressed in 0C.When the surface of the conductor is irregular.98 to 0. Typical values of this factor are m0 = 1. and hence by the air density. for cables of more than 7 strands.93 for roughened conductors.

386 p δ = . E0.rms = 21. .2 δ m0 r loge (d/r) kV to neutral 1.1 J R Lucas Breakdown of Gaseous Insulation p 273+ 20 0.2. For the formation of visual corona.6.4 Visual Corona Visual corona occurs at a higher voltage than the disruptive critical voltage. = 760 273 + t 273+ t The disruptive critical voltage can then be written as in the following equation. a certain amount of ionization. and the raising of an electron to an excited state are necessary.

but due to excitation.The production of light by discharge is not due to ionisation. and subsequent giving out of excess energy in the form of light and other electromagnetic waves.0 for smooth conductors. the disruptive critical voltage has to be multiplied by a factor dependant on the air density and the conductor radius. .3  d  = 21. To obtain the critical voltage for visual corona formation.2 mv δ r 1 +  . loge r δ r  1 J R Lucas Breakdown of Gaseous Insulation The values of the irregularity factor mv for visual corona is given by mv = 1. The empirical formula for the formation of visual corona is E v. Further the value of the irregularity factor is found to be different.rms 0.

3 = 0.82 for decided corona on stranded wires (all over the wire) J R Lucas Breakdown of Gaseous Insulation .72 for local corona on stranded wires (patches) = 0.

loge ) .  loge + r . Thus   E ∆ξ = ∆ d  r loge  r        1 J R Lucas Breakdown of Gaseous Insulation d   E ( 1 . . so that the effective radius becomes (r + Δr). just on the limit of corona formation. ∆ r = d 2 r d r   ( r loge d ) 2  ( r loge )   r r   .2. Assume that there is a thin layer Δr of ionised air around each conductor. ∆ r  -E d r -d    r ∆ξ =  .5 Stable Corona formation Consider two conductors. The change in electric stress Δξ due to this layer can be determined using differentiation. 2   .6.1.

718). For example.When loge > 1. and finally leads to flash-over. For normal transmission lines. the effective limiting value of d/r is about 15 and not e (=2. the effective increase in diameter lowers the electric stress and no further stress increase is formed. the above expression is negative. and this causes a further ionisation and a further increase in radius. the ratio d/r is very much greater than 15 and hence stable corona always occurs before flashover. and corona is stable. If on the other hand. then the effective increase in the diameter raises the electric stress. d/r < e. the ratio has a value d/r = 4/16x10-3 = 250 >> 15. 3 J R Lucas Breakdown of Gaseous Insulation . in a 132 kV line with a spacing of 4 m and a radius of conductor of 16 mm. In practice. i. d/r > e (=2.e.718) Under this condition.

The more important effect is the radio interference. ( E . 10 . but will not be of sufficient importance to have any appreciable effect on the voltage regulation.rms ) 2 .1. The power loss due to corona is proportional to the square of the difference between the line-to-neutral voltage of the line and the disruptive critical voltage of the line.E 0. It is given by the empirical formula Pc = 243 r . ( f + 25 ) .rms = disruptive critical voltage (kV) .6. This loss will have a small effect on the efficiency of the line.2. .6 Power Loss due to Corona The formation of corona is associated with a loss of power.5 kW/km/phas e δ d 1 J R Lucas Breakdown of Gaseous Insulation where E0.

72] The Air density correction factor δ is given by . The operating frequency is 50 Hz. [The irregularity factors may be taken as mo = 0. and the power loss in the line due to corona.3 f = frequency of supply (Hz) J R Lucas Breakdown of Gaseous Insulation E = Phase Voltage (line to neutral) (kV) For storm weather conditions. both under fair weather conditions as well as stormy weather conditions for a 100 km long 3 phase. the disruptive critical voltage is to be taken as 80% of disruptive critical voltage under fair weather conditions. mv = 0.83. the Visual Corona inception voltage.54 cm. 132 kV line consisting of conductors of diameter 1. Example Determine the Disruptive Critical Voltage. arranged in an equilateral triangle configuration with 3 m spacing. The temperature of the surroundings is 400 C and the pressure is 750 torr.

925 x 0.77 x loge (3/0.0077) x [ 1 + 0.3 kV.72 x 0.3/(0.2 δ m0 r loge (d/r) kV to neutral = 21. = 0.2 x 0.925 x 0.386 p δ = .76 kV 132/3 = 76.895 x 0. = 760 273+ t 273+ t 750 293 δ = .3 J R Lucas Breakdown of Gaseous Insulation p 273+ 20 0.0077) = 74.2 x 0. Thus corona inception will normally occur.77)½] = 88.21 kV. Thus visual corona will normally not occur.83 x 0. Visual corona voltage = 21. Similarly. .77 x loge (3/0.925 760 313 ∴ Disruptive critical voltage = 21.

5 .8  . 10. Pc = 0.8 = 59. − 74.0077 .76*0.925 3  132  .1 kW/phase    3  2 1 J R Lucas Breakdown of Gaseous Insulation Similarly. and Visual corona inception voltage = 88. (50 + 25) . it would occur under stormy conditions.5 .8 = 70. 10.76  . 100 kW/phase = 2. 0.76 * 0.6 kV It is seen that although visual corona does not occur under normal conditions. power loss under stormy weather condition is given by 243 0. .8 kV. (50 + 25) .Power loss under fair weather condition is given by Pc = 243 0.925 3  132  .3*0.0077 . 100 kW/phase   3   = 268 kW/phase 2 Also corona inception voltage = 74. − 74.

2 Breakdown of Liquid and Solid Insulation .

.1Breakdown in Liquids In highly purified liquid dielectrics. The line up of particles is a fairly slow process. Unfortunately.3 Breakdown of Liquid and Solid Insulation J R Lucas 2. and is unlikely to affect the strength on voltages lasting for less than 1 ms. and distort the field so that breakdown occurs at relatively low voltage. Effect of these impurities is relatively small for short duration pulses (10 μs). forming a bubble. liquids are easily contaminated. Under the action of the electric field. the solid impurities line up at right angles to equipotentials. breakdown is controlled by phenomena similar to those for gasses and electric strength is high (of order of 1 MV/cm). However. if the voltage is applied continuously. and may contain solids. other liquids in suspension and dissolved gasses. dissolved gasses may come out of solution.

which can be made to act as barriers. preventing the line-up of solid impurities and localising of any bubbles which may form. they are used at much higher tresses (up to 1 MV/cm) in conjunction with solids. of the tendency to become contaminated. However.4 Breakdown of Liquid and Solid Insulation J R Lucas The gas in the bubble has a lower strength than the liquid. . liquids are not usually used alone above 100 kV/cm in continuously energised equipment. so that more gas is produced and the bubble grows. Because. ultimately causing breakdown. The main function of the liquid in such arrangements is to fill up the voids.

(c) the evolution of bubbles of gas and the formation of solid products of decomposition (if the liquid is of requisite chemical nature) (d) formation of small pits on the electrodes. (b) a bright luminous path from electrode to electrode.1 Breakdown of Commercial liquids When a difference of potential is applied to a pair of electrodes immersed in an insulating liquid. . The passage of a spark through a liquid involves the following. If the voltage is raised continuously.5 Breakdown of Liquid and Solid Insulation J R Lucas 2. (a) flow of a relatively large quantity of electricity. determined by the characteristics of the circuit. at a critical voltage a spark passes between the electrodes.1. (e) an impulsive pressure through the liquid with an accompanying explosive sound. a small conduction current is first observed.

(c) breakdown strength is independent of hydrostatic pressure for degassed oil. These impurities which lead to the breakdown of commercial liquids below their intrinsic strength. (b) breakdown strength decreases with increase in electrode spacing. can be divided into the following 3 categories. but increases with pressure if oil contains gases like nitrogen or oxygen in solution. the breakdown strength will depend more upon the nature of impurities it contains than upon the nature of the liquid itself. which may not be subjected to very elaborate purifying treatment. In the case of commercial insulating liquid.6 Breakdown of Liquid and Solid Insulation J R Lucas Tests on highly purified transformer oil show that (a) breakdown strength has a small but definite dependence on electrode material. .

. These will be considered in turn in the following sections. (c) Impurities which result in local enhancement of electric field in a liquid (ex: conducting particles). Breakdown of the impurities may trigger off the total breakdown of the liquid. (b) Impurities which are unstable in the electric field (ex: globules of water).7 Breakdown of Liquid and Solid Insulation J R Lucas (a) Impurities which have a breakdown strength lower than that of the liquid itself (ex: bubbles of gas). Instability of the impurity can result in a low resistance bridge across the electrodes and in total breakdown. The enhanced field may cause local breakdown and therefore initiate complete breakdown.

The elongation continues.8 Breakdown of Liquid and Solid Insulation J R Lucas 2. or other causes.1.2 Breakdown due to gaseous inclusions Gas or vapour bubbles may exist in impure liquid dielectrics. when sufficient electric field is applied. temperature and pressure variations. and at a critical length the gas inside the bubble (which has a lower breakdown strength) . The electric field Eb in a gas bubble which is immersed in a liquid of permittivity ε1 is given by Eb = 3ε1 E0 2 ε1+1 where E0 is the field in the liquid in the absence of the bubble. either formed from dissolved gasses. The electrostatic forces on the bubble cause it to get elongated in the direction of the electric field.

Consider a spherical globule of liquid of permittivity ε2 immersed in a liquid dielectric of permittivity ε1. an elongated spheroid). . the field inside the globule would be given by E = 3 ε1 E0 2 ε1+ε 2 where E0 is the field in the liquid in the absence of the globule. This discharge causes decomposition of the liquid molecules and leads to total breakdown. When it is subjected to an electric field between parallel electrodes.3 Breakdown due to liquid globules If an insulating liquid contains in suspension a globule of another liquid.e.1. The electrostatic forces cause the globule to elongate and take the shape of a prolate spheroid (i. then breakdown can result from instability of the globule in the electric field.9 Breakdown of Liquid and Solid Insulation J R Lucas breaks down. 2.

10 Breakdown of Liquid and Solid Insulation J R Lucas As the field is increased.Variation of ratio of ratio ε 2/ε stress E diameters of spheroid < 20 > 0. Ec / γ Electric Figure 2. the globule elongates so that the ratio γ of the longer to the shorter diameter of the spheroid increases.5 ε = of rit 1 = 1 diamete ∞ of rs spheroi d .1 . For the same field E. the ratio γ is a function of ε2/ε1.

1 . no stable shape exists. Example 2. and the globule keeps on elongating eventually causing bridging of the electrodes. and is given by Ecrit.11 Breakdown of Liquid and Solid Insulation J R Lucas When ε2 >> ε1 (generally when ε2/ε1 > 20). and the field exceeds a critical value.  σ = 1. When ε2/ε1 >> 20.542  R  ε1     1 2 E crit kV/cm where σ = surface tension of the globule (N/m) ε1 = relative permittivity of the insulating liquid R = initial radius of globule (m). the critical field at which the globule becomes unstable no longer depends on the ratio. and breakdown of the gap.

Thus Ecrit = 1. In fact. may greatly influence breakdown conditions. = 0. ε2 = 90) in an insulating oil (ε1 = 2). a globule of water of radius of only 0.043/10–6 x 2)½ kV/cm = 226 kV/cm. such as condensed breakdown products. which is quite unobservable. .043 N/m. or hygroscopic solid impurities.542 (0. and then electrode gap breakdown channels develop at the end of the globule. Thus even submicroscopic sources of water. ε2 >> ε1 . will be disrupted at a value of about 1 MV/cm which is the breakdown strength of the pure liquid.12 Breakdown of Liquid and Solid Insulation J R Lucas For a droplet of water (R = 1 μm . Also σ = 0. A globule which is unstable at an applied value of field elongates rapidly.05 μm.226 MV/cm Thus we see that a droplet of water even as small as 1 μm in radius (quite unobservable) can greatly reduce the breakdown strength of the liquid dielectric.

13 Breakdown of Liquid and Solid Insulation J R Lucas Propagation of the channels result in total breakdown. .

Generally ε2 > ε1 .14 Breakdown of Liquid and Solid Insulation J R Lucas 2.1. Particles will continue to move in this way and will line up in the direction of the field. so that the force would move the particle towards the regions of stronger field. solid impurities cannot be avoided and will be present as fibres or as dispersed solid particles. r = radius of particle. If the impurity is considered to be a spherical particle of permittivity ε2 and is present in a liquid dielectric of permittivity ε1.ε1 ) F = r ε0 ∆ E2 2 ε2 + 2 ε1 where E = applied field.4 Breakdown due to solid particles In commercial liquids. . it will experience a force 1 3 ( ε2 .

However they may be used up to 1 MV/cm in conjunction with solids which can be made to act as barriers. . liquids are seldom used alone above 100 kV/cm in continuously energised equipment. Because of the tendency to become contaminated. preventing the line–up of solid impurities and localising bubbles which may form. which at a critical length may cause breakdown.15 Breakdown of Liquid and Solid Insulation J R Lucas A stable chain of particles would be produced.

16 Breakdown of Liquid and Solid Insulation J R Lucas 2. . This is done by distillation and degassing.5 Purification of a liquid for testing (a) Removal of dust Small dust particles can become charged and cause local stresses which can initiate breakdown. Thus it necessary to control the amount of gases present. but oxygen and carbon dioxide can cause the strength to change significantly.1. Careful filtration can remove dust particles greater in size than 1 μm. Some gases such as Nitrogen and Hydrogen do not appear to upset the electrical properties to a great extent. They can also coalesce to form conducting bridges between electrodes. (b) Removal of dissolved gasses Liquid insulation will normally contain dissolved gas in small but significant amounts. The strength of the liquid then increases and greater stability is achieved.

Water can be removed by drying agents. where test cells are small. . Not only is the surface smoothness important. electrode preparation is much more critical than it is for measurements on gases or solids. particularly oxides can have a marked influence on the strength. but surface films. For measurements on liquid dielectrics. and by freezing out in low temperature distillation.17 Breakdown of Liquid and Solid Insulation J R Lucas (c) Removal of ionic impurities Ionic impurities in the liquid (particularly residual water which easily dissociates) leads to abnormal conductivity and heating of the liquid. vacuum drying.

(a) electro–mechanical breakdown (b) breakdown due to internal discharges (c) surface breakdown (tracking and erosion) (d) thermal breakdown (e) electro–chemical breakdown (f) chemical deterioration These will now be considered in the following sections. Dielectrics usually fail at stresses well below the intrinsic strength due usually to one of the following causes.18 Breakdown of Liquid and Solid Insulation J R Lucas 2. highly purified and free of imperfections. the breakdown strength is high.2Breakdown of Solid Insulating Materials In solid dielectrics. of the order of 10 MV/cm. . The highest breakdown strength obtained under carefully controlled conditions is known as the "intrinsic strength" of the dielectric.

a mechanical force will be exerted on the dielectric due to the force of attraction between the surface charges. d absence of field Figure 2.19 Breakdown of Liquid and Solid Insulation J R Lucas 2.1 Electro–mechanical breakdown– When an electric field is applied to a dielectric between two electrodes.2. This is shown in figure 2. This compression decreases the dielectric thickness thus increasing the effective stress.2 .Process of breakdown with applied field o .2.

Pc = Y ln (do/d) At equilibrium.6 do. equating forces gives the equation.20 Breakdown of Liquid and Solid Insulation J R Lucas Compressive force Pc = ½ D E = ½ εo εr V2/d2. Thus when the field is increased. At the field when d < 0. the thickness of the material decreases. any further increase in the field would cause the mechanical collapse of the dielectric. The apparent stress (V/do) at which this collapse occurs is thus given by the equation . and From Hooke's Law for large strains.6 do. it is seen that the system becomes unstable when ln (do/d) > ½ or d < 0. V = 2 2Y εo εr d ln 2 do d By differentiating with respect to d.

6   ε o ε r  1 2 .21 Breakdown of Liquid and Solid Insulation J R Lucas  Y  E a = 0.

u .3. the field in the voids may exceed utheir breakdown value and breakdown may occur. The voltage across the E qvoid again builds up and the discharges recur. When the voltage Vv across the void exceeds the critical voltage Vc. These voids have a dielectric constant of unity and a lower dielectric strength.22 Breakdown of Liquid and Solid Insulation J R Lucas 2. a discharge 3is initiated and the voltage collapses. The discharge extinguishes very rapidly (say 0.2. . shown in figure 2. C V vs ≡ Hence the electric field strength in the voids is higher than that across the F oid dielectric. r eThe mechanism can be explained by considering the following equivalent circuit 2of the dielectric with the void.1 μs). p iv gThus even under normal working voltages.2 Breakdown due to internal discharges Solid insulating materials sometimes contain voids or cavities in the medium or boundaries between the dielectric and the electrodes.

v a 23 l eThe number and frequency of the discharges will depend on the applied voltage. f d i e l e c tr i c Breakdown of Liquid and Solid Insulation J R Lucas . ir each of the discharges. u The gradual erosion of the material and consequent reduction in the thickness of it othe insulating material eventually leads to breakdown.4. n tThe voltage and current waveforms (exaggerated for clarity) are shown in figure c2. there will be heat dissipated in the voids which will In ccause carbonization of the surface of the voids and erosion of the material.

Deterioration due to internal discharges .it h v24 o i d Breakdown of Liquid and Solid Insulation J R Lucas i t – V v igure applied 2.4 voltage across void if no F VC voltage discharges discharges occurred Internal C Breakdown by this process is slow and may occur in a few days or may take a few years.

25 Breakdown of Liquid and Solid Insulation J R Lucas In organic liquid-solid dielectrics. All voids in the dielectric can be removed by careful impregnation and this results in an increase in the discharge inception stress Ei. In oil impregnated paper these are (a) decomposition of moisture in paper (b) local electrical breakdown of the oil. internal discharges produce gradual deterioration because of (a) disintegration of the solid dielectric under the bombardment of electrons set free by the discharges (b) chemical action on the dielectric of the products of ionization of the gas (c) high temperatures in the region of the discharges. . The final value Ei then depends on electrical processes which lead to gas formation.

resulting in further gas formation and a rapid growth of the bubble. Also. widespread carbonization occur. permanent damage has been caused by the discharges and this manifests itself in an increase of loss angle and is due to the formation of ions by the discharges. discharges in the bubble decompose the molecules of the oil. Although on resting Ei improves. often lower than the rated stress. due to the discharges. the gas first formed arises from electrochemical decomposition of water held in the paper. Except in very dry conditions. but resting the dielectric long enough for the gas to dissolve in the oil restores the initial high discharge inception stress. the inception stress Ei is low. As long as the bubble remains in the dielectric.26 Breakdown of Liquid and Solid Insulation J R Lucas Stress at which gas is evolved from paper containing appreciable quantities of moisture can be less than 10 V/μm. When a gas bubble is formed in an oil-paper dielectric at the discharge inception stress Ei. but increases continuously with increasing dryness and can be higher than 100 V/μm when the paper is thoroughly dry. .

3 Surface Breakdown Surface flashover Surface flashover is a breakdown of the medium in which the solid is immersed. and the stress can reach up to εr times the mean stress in the gap. In particular if the edges are chipped. then the voltage gradient is not affected by the solid insulation. If a piece of solid insulation is inserted in a gas so that the solid surface is perpendicular to the equipotentials at all points.27 Breakdown of Liquid and Solid Insulation J R Lucas 2.2. . An example of this is a cylindrical insulator placed in the direction of a uniform field. then an air gap exists next to the electrode. The role of the solid dielectric is only to distort the field so that the electric strength of the gas is exceeded. Field intensification results if solid insulation departs even in detail from the cylindrical shape. or if the ends of the cylinder are not quite perpendicular to the axis.

The conducting film is usually moisture from the atmosphere absorbed by some form of contamination. Discharge may therefore occur at a voltage approaching 1/εr times the breakdown voltage in the absence of the cylinder.28 Breakdown of Liquid and Solid Insulation J R Lucas [εr is the dielectric constant of the cylinder]. (c) degradation of the insulation must be caused by the sparks. . Moisture is not essential as a conducting path can also arise from metal dust due to wear and tear of moving parts. The three essential components of the surface flashover phenomena are (a) the presence of a conducting film across the surface of the insulation (b) a mechanism whereby the leakage current through the conducting film is interrupted with the production of sparks. and these discharges can precipitate a breakdown.

Paschen minimum in air at N.29 Breakdown of Liquid and Solid Insulation J R Lucas Sparks are drawn between moisture films.] Degradation of the insulation is almost exclusively the result of heat from the sparks. For example. For a discharge to occur. and this heat either carbonises if tracking is to occur. It does not depend on gaseous breakdown. but in most practical cases. there must be a voltage at least equal to the Paschen minimum for the particular state of the gas. whereas tracking can occur at well below 100 V. or volatilises if erosion is to occur. Increase of creepage path during design will prevent tracking. separated by drying of the surface due to heating effect of leakage current. which act as extensions to the electrodes. moisture films can eliminate the designed creepage path.T. Carbonization results in a permanent extension of the electrodes and usually takes the form of a dendritic growth.P it is 380 V. .

The insulating material must be organic in nature for tracking to occur. the process is simply one of pitting. . If surface discharges are likely to occur. and in most cases the conduction (carbon path) results from degradation of the insulation itself leading to a bridge between the electrodes. if the products of decomposition are volatile and there is no residual conducting carbon on the surface. whereas erosion only weakens the material but allows operation until replacement can be made later. it is preferable to use materials with erosion properties rather than tracking properties. which again occurs in organic materials.30 Breakdown of Liquid and Solid Insulation J R Lucas Tracking Tracking is the formation of a permanent conducting path across a surface of the insulation. Erosion In a surface discharge. This is erosion. as tracking makes insulation immediately completely ineffective.

the temperature of the insulation increases. The power dissipated in the dielectric can be calculated as follows.4 Thermal Breakdown Heat is generated continuously in electrically stressed insulation by dielectric losses.31 Breakdown of Liquid and Solid Insulation J R Lucas 2. Uniform direct stress Power dissipated/volume = ξ2/ρ W/m3 where ξ = uniform direct stress V/m ρ = resistivity of insulation Ω m . If the heat generated exceeds the heat lost to the surroundings. which is transferred to the surrounding medium by conduction through the solid dielectric and by radiation from its outer surfaces.2.

tan δ . A . εr ε0 . d .32 Breakdown of Liquid and Solid Insulation J R Lucas Uniform alternating stress Power dissipated P = V . 2 π f . I cos φ where V = applied voltage V ω = supply frequency δ ε ξ = loss angle rad Hz C = dielectric capacitance F = A εr ε0 / d = dielectric constant = alternating stress V/m = V . VCω tan δ Therefore P = V2 (A εr ε0 / d) ω tan δ Re-arranging terms gives the result P = (V/d)2 .

556 ξ2 f εr tan δ W/m3 W/m3 if ξ is in kV/cm W/m3 with ξ in kV/cm = 2 π x 8.563 x 10-11 ξ2 εr f tan δ x 1010 Power dissipated/volume= 0. (i. where θ = ambient temperature .θ0).d is the volume of the dielectric.e. and V/d is the uniform applied stress. Heat lost = k (θ .854 x 10-12 ξ2 εr f tan δ W/m3 The simplest case is where the loss of heat by cooling is linearly related to the temperature rise above surroundings. and the heat generated is independent of temperature. the resistivity and the loss angle do not vary with temperature).33 Breakdown of Liquid and Solid Insulation J R Lucas Since A. Power dissipated/volume= ξ2 εr ε0 2 π f tan δ = 5.

.5 lost generat ea rature Thermal 1 0 ed t breakdown Equilibrium will be reached at a temperature θ1 where the heat generated is equal to the heat lost to the surroundings.34 Breakdown of Liquid and Solid Insulation J R Lucas heat h tempe θ Figure 2. as shown in figure 2.5.

Figure 2.35 Breakdown of Liquid and Solid Insulation J R Lucas E heat h tempe θ Figure 2. . although the heat lost may be considered somewhat linear.6 Thermal breakdown lost generat ea rature with non-linear heat generation 2 1 B A 0 ed t In practice.6 shows the variation of heat generated by a device for 2 different applied fields and the heat lost from the device with temperature. the heat generated increases rapidly with temperature. and at certain values of electric field no stable state exists where the heat lost is equal to the heat generated so that the material breaks down thermally.

36 Breakdown of Liquid and Solid Insulation J R Lucas The rapid increase is due to the fact that with rise in temperature. For the field E1. a stable temperature θA exists (provided the temperature is not allowed to reach θB). . The maximum voltage a given insulating material can withstand cannot be increased indefinitely simply by increasing its thickness. Usually. where T is the absolute temperature). This is because with thick insulation. the loss angle of the dielectric increases in accordance with an exponential law (loss ∝ e-A/T. beyond which it is not possible to go without thermal instability. Vθ is a limiting factor only for high-temperature operation. the heat generated is always greater than the heat lost so that the temperature would keep increasing until breakdown occurs. For the field E2. Owing to thermal effects. or at high frequency failures. there is an upper limit of voltage Vθ. the internal temperature is little affected by the surface conditions. in the practical use of insulating materials.

5 Electro–chemical Breakdown– Since no insulant is completely free of ions. The ions may arise from dissociation of impurities or from slight ionisations of the insulating material itself. . The reactions are much slower than in normal electrolytic processes due to the much smaller currents. When these ions reach the electrodes. reactions occur in accordance with Faraday's law of electrolysis. The insulation and the electrode metal may be attacked. gas may be evolved or substance may be deposited on the electrodes. The products of the electrode reaction may be chemically or electrically harmful and in some cases can lead to rapid failure of the insulation.2.37 Breakdown of Liquid and Solid Insulation J R Lucas 2. but on a much smaller scale. a leakage current will flow when an electric field is applied.

care must be taken to prevent contamination during manufacture. Typically a 1 μF paper capacitor operating at 1 kV at room temperature would require 2 to 3 years to generate 1 cm3 hydrogen.38 Breakdown of Liquid and Solid Insulation J R Lucas The products of the reactions may be electrically and chemically harmful because the insulation and electrodes may be attacked. . and because harmful gases may be evolved. the products of electrolysis would be formed much more rapidly. This is due to the fact that the reactions may be wholly or partially reversed when the polarity changes and the extent of reaction depends on the reaction rate and the time for diffusion of the reaction products away from the electrodes as well as on the nature of the reaction products. At elevated temperatures. Also since impurities give rise to an increase in the ion concentration. The rate of electrolysis is much greater with direct stress than with alternating stress.

Hydrogen ions discharged at the cathode readily react with the stabilizer rather than with the impregnant. a more difficult chemical process. The long term lives of capacitors containing chlorinated impregnants under direct stress may be greatly extended by adding small quantities of certain stabilizers.39 Breakdown of Liquid and Solid Insulation J R Lucas However at power frequency. The most frequent source of ions is ionizable impurities in the insulation. Thus contamination of insulation during manufacture and during assembly into equipment must be avoided with great care. Also. which are hydrogen acceptors and act as depolarizers at the cathode. contamination in polar insulating materials should be avoided with still greater care because of the greater degree of dissociation of ionic substance in solution. electrochemical effects can be serious and are often responsible for long-term failure of insulation. .

and rapid deterioration occurs due to attack of the acid on the electrodes and cellulose.40 Breakdown of Liquid and Solid Insulation J R Lucas In the absence of the stabilizer. mean life may be extended 50 times. . with 2% of the stabilizer Azobenzene. The extension of the life caused by the stabilizers is proportional to the amount of stabilizer added. the hydrogen reacts with the chlorine of the impregnant to produce hydrochloric acid. For example.

this process is very slow. Such chemical changes may result from spontaneous breakdown of the structure of the material. The logarithm of the life t of paper insulation can be expressed as an inverse function of the absolute temperature θ. but the process is strongly temperature dependant.2. especially organic materials. . show chemical instability. Chemical Instability Many insulating materials.41 Breakdown of Liquid and Solid Insulation J R Lucas 2.6 Chemical Deterioration Progressive chemical degradation of insulating materials can occur in the absence of electric stress from a number of causes. Under normal operating conditions.

0.7 .1% o 0o (weeks 0 ) water ) where A & B are constants If t is expressed in weeks. for vacuum dried paper immersed in oil in contact with Nitrogen. (K 8 2 Life of Figure 2. the constants have values A = 7000 and B = .0 0. .16. 0 1o.42 Breakdown of Liquid and Solid Insulation J R Lucas log10 t = A/θ + B 0.Dependence of ( × − 1 028 026 024 022 02 01 1 0) of paper on temperature 5 0 Paper life C 0.

so that t decreases by a factor of about 6. Electrical properties of materials such as paper. B decreases so that the life of the paper also decreases.8. Hydrolysis When moisture or water vapour is present on the surface of a solid dielectric. Figure 2.1% moisture present. .7 shows the variation.1% moisture reduces the life of the insulation by as much as 6 times. and other cellulose materials deteriorate very rapidly due to hydrolysis. hydrolysis occurs and materials lose their electrical and mechanical properties. the life of the insulation decreases much more rapidly.43 Breakdown of Liquid and Solid Insulation J R Lucas In the presence of oxygen or moisture. cotton tape. With about 0. With increase in amount of moisture present. B decreases by as much as 0. This means that presence of about 0.

and cellulose degrades much more rapidly in the presence of traces of acidic substances). materials such as rubber and polyethylene undergo oxidation giving rise to surface cracks. Other processes Progressive chemical degradation of insulating materials can also occur due to a variety of processes such as. and . Oxidation In the presence of air or oxygen. incompatibility of materials (ex: rubber ages more rapidly at elevated temperatures in the presence of copper.44 Breakdown of Liquid and Solid Insulation J R Lucas Polyethylene film may lose its mechanical strength in a few days if kept at 100 % relative humidity. particularly if stretched and exposed to light. Polythene also oxidises in strong day light unless protected by an opaque filler. especially ozone.

either in series. parallel or both. composite dielectrics occur either due to design considerations (ex: paper with an impregnating liquid) or due to practical difficulties of fabrication (ex: air in parallel with solid insulation). 2.45 Breakdown of Liquid and Solid Insulation J R Lucas leaching (washing out of a soluble constituent) of chemically active substances (ex: glass fabrics made from glasses of high sodium content lose their strength rapidly due to leaching of sodium to the surface of the fibres and the subsequent chemical attack of the strong alkali on the glass surface). Usually more than one insulating material will be involved. The simplest form of composite insulation system consists of 2 layers of the same material. . In this case advantage is taken of the fact that two thin sheets have a higher electric strength than a single sheet of the same total thickness.3Breakdown of Composite Insulation Almost no complete electrical insulation consists of one insulating phase. In other cases.

(iii) the internal or partial discharge products of one component invariably affect the other components in the system. (i) The stress distribution at different parts of the insulation system is distorted due to the component dielectric constants and conductivities. the system as whole has to be considered. The following considerations determine the performance of the system as a whole. . (ii) the breakdown characteristics at the surface are affected by the insulation boundaries of various components. the behaviour of the composite insulation could be predicted from the behaviour of the components.46 Breakdown of Liquid and Solid Insulation J R Lucas In certain cases. and (iv) the chemical ageing products of one component also affect the performance of other components in the system. But in most cases.

47
Breakdown of Liquid and Solid Insulation J R Lucas

48
Breakdown of Liquid and Solid Insulation J R Lucas

2.3.1 Matching dielectric constants When composite insulation has components with different dielectric constants, utilisation of the materials may be impaired. This is especially true in the oil/transformerboard dielectric. This is because the oil has a lower dielectric constant and lower dielectric strength compared to that of transformerboard. Since the dielectrics are in series,

V ε2 .V ξ1 = 1 = d1 ε1 d 2 +ε 2 d1

49
Breakdown of Liquid and Solid Insulation J R Lucas

V ε1 .V ξ2 = 2 = d2 ε1 d 2 +ε 2 d1

Figure 2.8 Composite Dielectric
5 d1 V1// 0 2 2 0

50
Breakdown of Liquid and Solid Insulation J R Lucas

V 1 = C2 = A ε 2 . d1 = ε 2 d1 V2 C1 d2 Aε1 ε1d2 V = V 1 +V 2 V1 = V1 = ε 2 d1 V V 1 +V 2 ε1 d2+ε 2 d1

Example A transformer oil having a dielectric constant of 2.2 and a dielectric strength of 25 kV/mm, is used as an insulation in a of spacing 8 mm. Determine the maximum applicable voltage. A barrier of thickness 3 mm of transfomerboard with a higher dielectric strength of 50 kV/mm (dielectric constant 4.4) is used in this space to increase the strength. Does the transformerboard serve this purpose in this case ? With only transformer oil, the maximum applicable voltage is given by

51
Breakdown of Liquid and Solid Insulation J R Lucas

V = 25× 8 = 800 kV If a barrier of thickness 3 mm is placed in the space with the oil, the maximum applicable voltage is given by
50 = 4.4 ×V 2.2 × 2 + 4.4 × 6 50 × 14 V = = 350 kV 2

It can be seen that the maximum applicable voltage in fact reduces below that of only oil. It is thus important, when barriers have to be used, to match the permittivities of the component insulations. Thus great gains could be achieved if a transformerboard with a dielectric constant of 2.6 could be used instead of one with 4.4.

52
Breakdown of Liquid and Solid Insulation J R Lucas

3
High Voltage Cables

54

J.R.Lucas

High Voltage Cables

3.0High Voltage Cables
High Voltage Cables are used when underground transmission is required. These cables are laid in ducts or may be buried in the ground. Unlike in overhead lines, air does not form part of the insulation, and the conductor must be completely insulated. Thus cables are much more costly than overhead lines. Also, unlike for overhead lines where tappings can easily given, cables must be connected through cable boxes which provide the necessary insulation for the joint.
Cables have a much lower inductance than overhead lines due to the lower spacing between conductor and earth, but have a correspondingly higher capacitance, and hence a much higher charging current. High voltage cables are generally single cored, and hence have their separate insulation and mechanical protection by sheaths. In the older paper insulated cables, the sheath was of extruded lead.

55

J.R.Lucas

High Voltage Cables

Figure 3.1 shows three such cables, as usually laid out.
Figure 3.1 - Layout of three, single-core cables

The presence of the sheath introduces certain difficulties as currents are induced in the sheath as well. This is due to fact that the sheaths of the conductors cross the magnetic fields set up by the conductor currents. At all points along the cable, the magnetic field is not the same. Hence different voltages are induced at different points on the sheath. This causes eddy currents to flow in the sheaths. These eddy currents depend mainly on (a) the frequency of operation, (b) the distance between cables, (c) the mean radius of the sheath, and

56

J.R.Lucas

High Voltage Cables

(d) the resistivity of the sheath material.

3.1Power loss in the Cable
Heat dielectri 3.2 - Heat Transfer conducto sheath Figure generate c loss & r in a cable due to losses loss d interheat h loss

Power loss in the cable can occur due to a variety of reasons (Figure 3.2). They may be caused by the conductor loss (also sometimes called the copper loss on account of the fact that conductors were mainly made out of copper) due to the conductor current passing through the resistance of the conductor, dielectric losses caused by the voltage across the insulation,

R.57 J. The dielectric loss is voltage dependant.Lucas High Voltage Cables sheath losses caused by the induced currents in the sheath. . while the rest is current dependant. and intersheath losses caused by circulating currents in loops formed between sheaths of different phases.

58 J.Lucas High Voltage Cables 3. and hence there is a power loss. then .1. The current leads the voltage by an angle of less than 90o. the power factor is not zero. the power factor is zero. Since the cable is not a perfect dielectric. and E is the applied voltage.3 Loss E angle φ δ For a perfect dielectric. If C is the capacitance of the cable.R.1 Dielectric loss IFigure 3.

1.1. 3.2 Conductor loss The conductor loss Pc is given by Pc = I2 Rc watt where Rc is resistance of the conductor and I is the current in the cable.59 J. 3.Lucas High Voltage Cables charging current Ic = E C ω power loss P = E I cos φ = E I sin δ = E (I/cos δ) sin δ = E2 C ω tan δ The power loss is proportional to E2 and tan δ .3 Sheath loss Losses occurring in the sheath of a cable is usually obtained by the empirical Arnold's formula Psh given by .R.

4 Intersheath Loss Intersheath losses are caused by the induced emf between the sheaths causing a circulating current. 3.R.7 × 10-3 I  r m  watt Psh Rsh  d  where rm = mean radius of sheath d = distance between cables (centre to centre) Rsh = resistance of full length of cable I = current in cable The sheath loss is usually about 2 to 5 % of the conductor loss.60 J.Lucas High Voltage Cables   = 7. as otherwise sparking could occur causing damage to the sheaths.1. This loss is thus present only when the sheaths of adjacent cables are connected together. 2 2 . The sheaths need to be connected together in practice.

61 J.R. Msh and the induced current Iish is given by I ish = [R E ish 2 sh + ω M sh 2 2 ] [R 1 2 = i ω M sh 2 sh + ω M sh 2 2 ] 1 2 Therefore the intersheath loss Pish is given by . ω . The mutual inductance Msh between a core of one cable and the sheath of an adjacent cable is given by M sh = µ d  ln   2π  r  The voltage induced Eish is given by Eish = I .Lucas High Voltage Cables The intersheath loss Pish can be calculated in the following manner.

R sh Pish + ω 2 M sh R sh 2 2 2 Generally.7 x 10  r m  I 2 ω 2 M sh 2 2 2   + P loss = I R + I d  R sh R sh -3 2 .R.62 J. Thus the total power loss (exclusive of the dielectric loss) is given as Total Power loss = Pc + Psh + Pish 7.Lucas High Voltage Cables ω = I ish2 R sh = I 2 M sh . the sheath resistance Rsh >> ω Msh so that =I Pish 2 ω M sh R sh 2 2 The intersheath loss is larger than the sheath loss and may range from 10% to 50% of the copper loss.

5 Cross-bonding of Cables .1.7 x 10  r m  ω 2 M sh2   + Reff = Rc + d  R sh R sh -3 2 Since the sheath loss is usually very small.63 J.R.Lucas High Voltage Cables Since the whole expression is dependant on I2. the effective conductor 2 resistance can be written 2 ω M sh as R eff = Rc + R sh 3. This gives the total power loss in terms of the effective resistance as Ptotal = I2 Reff 7. we may express the loss in terms of an effective resistance Reff.

.4).R. currents are induced in the sheaths and lead to sheath circulating currents and power loss. These may be substantially reduced. and the current rating of the cable increased by cross bonding of the cables (Figure 3. Cross bonding of cables are done except for very short lengths of cable.Cross bonding of sheaths a c b When three single phase cables are used in power transmission.64 J.Lucas High Voltage Cables Figure 3. the cables between two adjacent discontinuities is a minor section.4 . The continuity of each cable sheath is broken at regular intervals.

The cable is also transposed (Figure 3. these points are earthed through non-linear resistors which limit voltage build up. To prevent excessive voltage build up at the cross bonded points.5) . A residual sheath voltage exists. especially during faults.Lucas High Voltage Cables Figure 3.R. where the sheaths are solidly bonded together and to earth. and the desired balance. giving negligible sheath voltage between the solid grounded positions is achieved by transposing the cables at each cross-bonded section.65 J.5 – Non-linear resistor earthing Three minor sections make up a major section.

The impregnating compound varies from manufacturer to manufacturer. . The paper must be free from ligneous fibres and from metallic or other conducting spots.2Impregnated Paper Insulation The insulation consists mainly of paper tape impregnated with compound. The compound with which the paper is insulated should be of such a consistency that it is plastic at ordinary temperatures. The paper is made from Manila fibre or wood pulp. with resin frequently added to lower the viscosity and to improve its impregnating qualities. and has no tendency to drain away from the cable.R. but they all are based o paraffinic or naphthenic mineral oil.Lucas High Voltage Cables 3.66 J.

the weakest part of the dielectric a breaksdown and deterioration starts getting more and more. t ( 3 k p . t(15 0 g d h u .R. e B The cause of breakdown is usually the non-homogeneity of the dielectric. be s The dielectric strength of impregnated paper is about 200 to 300 kV/cm. they may be able to withstand about 400 to 600 kV/cm. k d o w .67 J.atmospher o re) v u e . u r Initially.Lucas High Voltage Cables time 1 2 3 4 w 5 b F nder U 0 (hrs) 0 ipressure . V r 6 ) e Impregnated paper can withstand an electric stress if about 5 to 10 times that which could s withstood by dry paper insulation. r e When a test voltage is applied.

t The deteriorationaresults in the formation of voids and gasses. h Thus long term a breakdown strength and the instantaneous break down strengths differ. the breakdown stress will r continue to decrease and ultimately leads to breakdown. i Using a safety factor. r a This value may decrease with time due to deterioration to about 160 to c 200 kV/cm. g When the voltage is raised.Lucas High Voltage Cables o This is accentuated by the fact that the cable is not carrying the same l current all the time. s t i c o . c Some oil suffers condensation and hydrogen and other gases are evolved. not more than about 40 kV/cm is allowed in service.n v J. t In the case of a e badly impregnated dielectric. ionisation or glow discharge can occur in the e 68 voids and ionic bombardment of thee surface.R.

n 3. 1.2.R. High Insulation resistance i High dielectric strength n s Good mechanical strength u Immune to attack by acids and alkali in the range 0 . 3.Lucas 3.100o C l a Should not be too costly t Should not be hygroscopic (tending to absorb water). i o or if hygroscopic should be enclosed in a water tight covering. .p 69 a J. 2. 5.2. it has been observed that the long term breakdown stress is increased if the cable is subjected to pressure. This is due to the fact that the pressure discourages the formation of voids.2 Principle underlying the design of high voltage cable insulation By means of dielectric tests on cables.1 Propertiesprequired of cable insulation e Dielectrics used for cable insulation must have the r High Voltage Cables following properties. 4. 6.

Lucas High Voltage Cables Even for a badly impregnated cable.0 time tan impregnated (a) ( 1 atmos) (hrs) ) stress m 0 1 δ 15 atmospheres (a) (kV/cm) (b) badly impregnated ( 1 atmos) (b) 8 atmospheres (c) badly impregnated (15 atmos) (c) 1 atmosphere .70 J.R. ( 20 30 40 50 breakdown 100 (c (b (a applied kV/c 10 8 6 4 2 0a 0. If the cable is subjected to a pressure of about 15 atmospheres.7). the long term dielectric strength improves to about 400 kV/cm and a working stress of about 150 kV/cm may be used (Figure 3. the application of pressure improves the power factor (or loss tangent) considerably.

using of better materials. In modern high voltage cables.3 Paper insulated power cables .01) to about (0. 3. The paper is porous and contains in itself the impregnating compound. Further curves show how long term breakdown stress is improved by pressure. and (c) badly impregnated cable at a pressure of 15 atmospheres for about 47 hours. shows advantages of the pressure on the reduction of power factor. For high voltage cables. (b) badly impregnated cable at atmospheric pressure. impregnated paper insulation is very commonly used.Lucas High Voltage Cables Comparison of curves for (a) well impregnated cable at atmospheric pressure.003).2.71 J.007 to 0.R. power factor has been reduced from about (0. There are no voids present as the oil is present between the layers of the paper which forms the insulation.002 to 0.

pressurised cables are used where the impregnated paper insulation is kept under pressure. As paper is hygroscopic. a seamless lead sheath is extruded over the insulation so that no moisture will get in. and this is lapped round with the paper tape. and while still under vacuum.R. . Impregnating of the paper prevents void formation in the dielectric.i.72 J. (about 0. For high voltages. A vacuum is then applied for 20 to 50 hours to get rid of any trapped air inside the cable.s.35 MN/m2) is applied. impregnating compound is poured into the tank and thereafter a pressure of 50 p. It is first heated to about 100oC taking care not to burn it. as voids can easily lead to the breakdown of the dielectric.Lucas High Voltage Cables The conductor of the cable is stranded.

When the cable is pressurised. . With pressurised cables.R. The pressure may be applied by having either oil or gas under pressure. the long term breakdown strength does not differ much from the short term strength. a working stress of about 100 to 120 kV/cm may be used.Lucas High Voltage Cables A pressure of about 15 atmospheres is applied so that any potential voids would be instantaneously filled.73 J. longitudinal reinforcement to prevent bulging and reinforcement to prevent hoop stress are used. and as such using a safety factor.

R.Lucas High Voltage Cables 3. 8 C a b l e c .4 Insulation Resistance x R r Δ F ix g u r e 3 .74 J.2.

Lucas s 75 sFor a single core cable (figure 3. s ρ .665 R i .5 Capacitance in a single-core cable F i R .r o J. the insulation resistance between the -conductor and the outer sheath is given by the following.∆ x ∆ Res = e 2π x l c ti where l = length of cable(m) o ρ R dx ρ R ∴ ln( ) n Res = ∫ ∆ Res = ∫ = High Voltage Cables 2π l r x 2π l r 3.R.2.8). e . Res = ρ log10 ( ) x 10-10 M Ω l′ r where l′ = length of cable(km) 3.

q = charge/unit length of cable (C/m) x Δx C D = electric flux density r a = charge density (C/m2) b ε0 = permittivity of free space l = 1/(4π x 9 x 109) F/m e Consider an elemental cylinder of radius x and thickness dx.Lucas e Consider a single core cable (figure 3.u r 76 J. r (For impregnated paper insulation.9) with the following data.5) o High Voltage Cables s s s e . εr = 3. 9 R = radius of earthed sheath (m) . and of length c unity along the cable.R. 3 r = radius of core (m) .

V = q loge ( R/r ) = loge ( R/r ) 2π ε εr q εr ∴ capacitnac = = e F/m 9 V 18 x 10 loge ( R/r ) 0.c ti o n 77 J. ∴ D= dv also.e.0383ε r 0. dx 18 x 109 xq so that V = ∫ ξ dx = ∫ dx εr x r r R R q i .024 ε r ∴ C = µ F/mile = µ F/km log10 (R/r) log10 (R/r) 18 x 109 . ξ = .R.Lucas High Voltage Cables q 2π x D q 18 x 109 xq ∴ electricstress ξ = = = ε 2π ε0εr x εr x q = D x 2 π x x1 .

and its magnitude alternates with time. Aε ∆y.78 J. = constant. capacitance is constant. the stress is radial.l ε C= = d ∆x .Lucas High Voltage Cables Field Plotting In the case of the single core cables. Further for constant differences.R. equipot electric ∆ ential Vflux x y φ lines The electric flux lines and the equipotential lines are perpendicular to each other. ∆ φ (or  and ∆ V are constants. q) Thus for the elemental figure shown.

The electrical properties of paper varies in different directions. which are perpendicular to the equipotential lines. for a few points of the cycle are illustrated in figure 3. and tangential components of stress exist.R. In the case of the 3-core cable. the electric stress is not radial. the electrostatic field is a somewhat rotating field and not a pulsating one. If paper insulation is used around each cores.Lucas High Voltage Cables . Consequently. Thus curvilinear squares are formed in the sketch. 79 J. Typical variations of the equipotential surfaces. since the centres of the cores lie in a circle. are not radial to the individual cores.10. then tangential stresses will be applied along the surface of the paper rather than just across it. From these it will be seen that the field lines. y/∆ x = constant (usually chosen as 1 for convenience of drawing).e.i.

R. .Lucas High Voltage Cables The effective dielectric stress of paper insulation is much greater across the layers than along it. Thus the presence of tangential stress in paper insulation leads to greater risk of breakdown.80 J.

6 Three-core Cables When three phase power is being transmitted.2.R. either three single-core cables or a single three-core cable may be used. .81 J.Lucas High Voltage Cables 3.

R.Lucas High Voltage Cables electric stress lines equipotential lines 1.82 J.5 0 . C B A 0.

83 J.Lucas High Voltage Cables .R.

3 2 1 1 = 0.R.84 J.866 .Lucas High Voltage Cables V .866 1∠ 30o= 0 V2 V3 = – 0.

3 2 1 1 V1 1∠= 0.259 V2 V3 = – 0.Lucas High Voltage Cables .R.707 Figure 3.Equipotential lines in three-core cables .85 J.966 5o = – 0.10 .

steel wires are laid in one or two layers. the 3-cores are individually insulated with paper insulation. and the lead sheath is extruded over this. . there is generally bitumen to prevent damage.86 J. There are two types of armouring used for these cables.steel tape is usually wound in two layers with opposite directions of lay (ii) Steel wire armouring . Over the lead sheath.R. Belt insulation is used on top of all three core insulations. In buried cables. The filler spaces between the core insulation is also filled up with insulation. (i) Steel tape armouring .Lucas High Voltage Cables 3. additional protection is necessary to prevent damage. but depriving these of voids is much more difficult.7 Three-core belted type Cables In the case of a 3-core cable.2.

Lucas High Voltage Cables Capacitance of 3-core belted type F S c b h o e e n lt r a d e i ti u n h c n s ts u o u l rl a ti a ti o n o n .87 J.R.

88 J. The capacitance per unit length to neutral is given by If t = thickness of belt insulation T = thickness of conductor insulation d = diameter of conductor εr = dielectric constant C0 =  T +t   t t +1 log10  0. Simon's expression can be used to obtain this value.84     d   T  T        2 0.R.03 ε r µ F/km Measurement of capacitance of 3-core cables .52   .1.7   + 3.Lucas High Voltage Cables The capacitance between the conductor to neutral of 3-core belted cables (Figure 3.11) cannot be obtained by a simple derivation as for the single core cable.

Lucas High Voltage Cables C Figure 3.89 J.Cable Capacitances s .R.12 .

90 J. but can be lumped as shown in figure 3.Lucas High Voltage Cables In three-core cables. capacitance does not have a single value.12.R. (a) Strap the 3 cores together and measure the capacitance between this . C between each core and sheath = Cs C between cores = C These Capacitances can be separated from measurements as described in the following section.

91 J.Lucas High Voltage Cables (a) Strap the 3 cores together and measure the capacitance between this .R.

13 Cs Bri Capacitance dge measurement .R.Lucas High Voltage Cables Figure 3.92 J.

93 J.R. Measured value = Cm1 = 3 Cs Gives capacitance to the sheath as Cs = Cm1/3 (b) Connect 2 of the cores to the sheath and measure between the .13.Lucas High Voltage Cables (a) Strap the 3 cores together and measure the capacitance between this bundle and the sheath as shown in figure 3.

Lucas High Voltage Cables (b) Connect 2 of the cores to the sheath and measure between the .94 J.R.

Lucas High Voltage Cables ta Cs Figure 3.R.14 nBri Capacitance δdge measurement .95 J.

Lucas High Voltage Cables (b) Connect 2 of the cores to the sheath and measure between the remaining core and the sheath (Figure 3. C = (Cm2 .R.Cm1)/6 giving the capacitance between the conductors.e. Measured value Cm2 = 2 C + Cs i. Effective capacitance to neutral Co of any cores may be obtained by .Cs)/2 = (3 Cm2 .96 J.14).

97 J.Lucas High Voltage Cables Effective capacitance to neutral Co of any cores may be obtained by .R.

Lucas High Voltage Cables Figure 3.R.15 3 Cs Calculation of Co C .98 J.

it is generally observed that charring occurs at those places where the stress is tangential to the layers of paper. so that the cable in effect becomes 3 individual cables laid within the same protective covering. Can usually be accomplished by screening each core separately (or by having individual lead sheaths for each core).Lucas High Voltage Cables Effective capacitance to neutral Co of any cores may be obtained by considering the star equivalent (Figure 3.Cm 1 C0 2 6 In the breakdown of actual 3-core belted cables.15).99 J. Thus for the insulation to be effective. 1 3 Cm 2 .Cm 1 = C s + 3 C = C m 1+ 3 C0 3 6 3 1 = Cm 2 . .R. the tangential stresses in paper insulation should be preferably avoided.

R.100 J. there is no belt insulation.2. The screening of individual cores is generally thin and flexible so that there is not much power dissipation in them. .Lucas High Voltage Cables 3.8 Hochstadter or "H" type Cable conduct lead metallise copper paper 3.16 Figure or sheath d paper insulatio woven cable worming H-type n fabric In this type of cable (Figure 3.16).

These are used upto 66 kV. The screens are thin so that there is hardly any current induced. These are wrapped round with copper woven fabric (cotton tape into which are woven copper wire). the individual cores contain no lead covering. In the H-type cable. The cable has the additional advantage that the separation of the cores by thermal expansion or mechanical displacement cannot introduce stresses in the dielectric.Lucas High Voltage Cables All the individual screens are earthed so that the potential at these sheaths are all zero and thus the stress lines between the cores and screens would be now radial.101 J. The metallised screens help to dissipate the heat.R. The sheaths surrounding the insulation of the cores consist of metallised perforated paper. This outer screen is in contact with the inner screens and is earthed. .

2.9 S. then cable is armoured with steel wire and tape.Lucas High Voltage Cables The three cores are laid up with fillers in the ordinary way. 3. The wormings of the H-type cable are full of oil. If cable is to be buried.102 J.L. type Cable cor lead cotton core Figure 3.17 – S compou e type sheath tape cable insulatio Lnd jute n worming .R.

Lucas High Voltage Cables Another development of the screening principle is the SL type cable (Figure 3. The wormings of the filler spaces in the S. after being lapped with paper and cotton tapes are laid with tarred jute yarn to get a circular formation and then wrapped with hessian tapes to form a bedding for the armouring.L. each core is screened and then individually sheathed with lead or aluminium. The electrical and thermal advantages of H-type cables are also enjoyed by the S. . The three metal sheathed cores. These do not have an overall lead sheath.R. In this.17).L. type cables. type cables do not contain much oil as do not get any electric stress.103 J. The electric field in the insulation surrounding each core is naturally radial and the function of the screens in this case is to eliminate the possibility of any stress across the clearance space between core and sheath.

and appear to be of no practical significance.L. type construction is useful on short runs because the terminating equipment is simplified. Also the void formation in the filler spaces are of no consequence.104 J.R. but the resulting losses are small.L. cable are the seats of induced currents. The separate lead sheaths in the S.Lucas High Voltage Cables These cables are suitable for hilly routes. Also. as the absence of oil in the filler spaces lessens the risk of oil drainage. the S. .

R. insulation in cables occupies a greater portion of the cable space. Thus higher installation costs are involved.2.10 Copper Space Factor R r F Unlike in overhead lines.Lucas High Voltage Cables 3.105 J. .

Thus we define a space factor to indicate the utilisation of the space. Thus Space factor = r2/R2 t T r R F i g u r e 3 .Lucas High Voltage Cables Ideally we would like the insulation to occupy the minimum possible thickness.R. as this gives the minimum conductor perimeter for the greatest conductor area and given insulation thickness.18).106 J. the best space factor is obtained with a concentric arrangement (Figure 3. copper space factor = cross − section area of conductor cross − section area of whole cable For a single core cable.

1 J.R. c a b l F e i g High Voltage Cables .19).20). consisting of circular conductors within a circular sheath. t Space factor = 3 r2/R2 h rwhere T = thickness of core insulation. t = thickness of belt insulation e e and R1 = r + T -For the 3-core cable circular cross-section is not the best shape for the c conductors.Lucas 9 107 For the 3-core cable (Figure 3. o r Other shapes which gives better space factors are the elliptical shaped e conductors and the sector shaped conductors (Figure 3..

2 0 S p e c i a l s h a p e 108 J.R.u r e 3 .Lucas High Voltage Cables .

Since it is required that this maximum stress in the dielectric should be as o low g as possible.R. differentiating with respect to r for minimum ξmax gives i v e b High Voltage Cables . ξ x = u 2π ε r 2π ε x c v t so that ξ x = R o x loge r r sIt is seen that since x is the only variable. also. x = r).Lucas f 109 3.e. the maximum stress in the tdielectric occurs at the minimum value of the radius x (i.3Dielectric Stress in a Single Core Cable c The voltage difference across the conductor and the sheath of a single core o cable is given by n d q R q v= loge .s o J.

718 f r a c15 t Thus if the overall diameter of the cable is kept fixed.e tt J.   . then R/r = e is the o condition for minimum ξmax. loge + r . Since R/r = e.e. r sThis value of radius of conductor will generally be larger than would be required for current carrying capacity.R.e. the minimum value of ξmax is given by High Voltage Cables .Lucas e 110 r d ξ max s =0 dr p 2 a    V R  1  i.  . = e = 2. = 0 c r  r   r loge R/r   e R i .

111 J.Lucas High Voltage Cables V V = ξ max = r loge R/r r .R.

Lucas High Voltage Cables 0 r R ξ F x m .112 J.R.

21).113 J. The insulation is fully stressed only at the conductor. As can be seen (Figure 3. the dielectric is not equally stressed at all radii. . this value of radius may be achieved by using Aluminium or hollow conductors.Lucas High Voltage Cables Since the radius of the conductor that would be given from the above expression is larger than is necessary for current carrying capacity.R. in a cable of homogeneous insulation. and further away near the sheath the insulation is unnecessarily strong and thus needlessly expensive.

114 J. .3. and the conductor radius r.2 Capacitance Grading In this method of grading.R. ε2. respectively. Consider a cable graded by means of 3 layers of insulation. as shown in Figure 3. r2 and r3 = R respectively. (i) Capacitance grading (ii) Intersheath grading 3.3. having permittivities ε1.Lucas High Voltage Cables 3.22. the insulation material consists of various layers having different permittivities. Let the outer radii of these layers by r1.1 Cable Grading for Uniform Stress Distribution The electric stress in the dielectric may be more equally distributed by one of the two following methods. ε3.

R. 2 2 C a p a c it a n .Lucas High Voltage Cables 18 0 r R ξ F r1 m 3 2 1 i xr2 g u r e 3 .115 J.

V3 can be determined . Then z = ξ max r loge r 1 V1 r similarly V 2 . the dmaximum stresses in the layers are equated.Lucas r 116 High Voltage Cables aIn order to secure the same value of maximum stress in each layer.e G J.R. i q q q n = = g 2 π ε 0 ε 1 r 2 π ε 0 ε 2 r1 2 π ε 0 ε 3 r 2 ∴ ε 1 r = ε 2 r1 = ε 3 r 2 Let the voltage across the inner-most layer of insulation be V1.

0.r) ln  r r1 r2   R . without increasing the overall diameter of the cable.8 to 4. as paper insulation has permittivities limited to the range 2. r 2 > r r Hence by grading the insulation. the operating voltage can be raised.r) ln r 2 + ( r 2 .R.  r1 r2 r3  V = ξ max  r ln + r 1 ln + r 2 ln  r r1 r2   R R  = ξ max  r ln r 1 + r ln r 2 + r ln + ( r 1 .r) ln  r r1 r2 r1 r2   R R  = ξ max  r ln + ( r 1 .Lucas High Voltage Cables Therefore the total voltage across the dielectric can be obtained as follows. > ξ max r loge .117 J.r) ln r 2 + ( r 2 . A difficulty with this method is that we cannot obtain a wide range of permittivities in practice. since r1 > r.

it has been assumed that the maximum permissible stress is the same for all three dielectrics used. ξ2.R. and are ξ1. then the maximum stresses should be reached at the same time for the most economical operation of the insulation. If the maximum stress in the three sections are different.118 J. This condition gives us the result ξ 1 ε 1 r = ξ 2 ε 2 r1 = ξ 3 ε 3 r 2 . ξ3 respectively.Lucas High Voltage Cables In the above analysis.

Lucas High Voltage Cables 3.3.119 J. 2 3 – I n te rs h e at h G r .3 Intersheath Grading V 0 r R ξ2 F 1 r1 i xr 1 2 3 m 2 g u r e 3 .R.

These intersheaths are connected to tappings from the supply transformer. are the potential differences across the sections of insulation. then .. Since there is a definite potential difference between the inner and outer radii of each sheath. and the potentials are maintained at such values that each layer of insulation takes its proper share of the total voltage. but is divided into two or more layers by means of cylindrical screens or intersheaths (Figure 3. V2. If V1..R. V3. The intersheaths are relatively flimsy.Lucas High Voltage Cables method of grading. and are meant to carry only the charging current.23). . we can treat each section separately as a single core cable. the same insulating material is used throughout the cable.d i n 120 gIn this J..

.... r2 = r + 2d.Lucas High Voltage Cables ξ max = V1 r loge r 1 r = V2 r loge r 2 r1 = ... are in phase. . . Thus. . if V is the phase to neutral voltage.121 J. V3. rn = r + n d .R. .. and if r is the conductor radius. r1 = r + d. all potential differences V1. Since the cable insulation now consists of a number of capacitors in series. formed by the respective intersheaths. V2. .. + Vn In the particular case that all the n layers have the same thickness d. r3 = r + 3d.. we can also write V = V1 + V2 + V3 + .. .. . .. .

there is a considerable difficulty in arranging for many intersheaths.1) d Hence substituting for the different values of m..1) d ) loge r + (m .122 J.. = r+d r+2 d M r loge (r + d) loge r r+d n r+md where M = ∑ ( r + (m .R..1) d m =1 The voltage across the mth section is given by V r+md V m = [ r + (m ..Lucas High Voltage Cables ξ max = V V1 V2 = = . we can obtain the voltage across the various layers that have to be maintained to give equal maximum stress in each section. this difficulty being mainly associated with the provision of . In practice.1) d ] loge M r + (m .

Lucas High Voltage Cables the different voltages for the intersheaths. and the expression for the maximum stress then given by ξ max = V r 1 loge + r loge r 1 r r1 R For the purpose of comparison with the ungraded cable. Since R and r are both kept fixed. . and the expression for stress must be differentiated with respect to r1 to obtain the condition for the minimum value of the maximum stress. and as a result it is usual to design a cable of this type with only one intersheath.e. r1 is the only variable. with R/r = e). and introduce an intersheath at a radius r1. This simplifies the design calculations.123 J. let us first take the optimally designed ungraded cable (i.R.

76 r  1. .R.76 r.76 r loge   + r loge    1.33 r  er   1.R  2 so that r1 .  . R  r1   R r 1  + loge + r .1 + loge R r1 + r r1 =0 R considerin g = e.124 J. also loge e = 1 r r r ∴ 1 loge 1 = 1 r r ∂ξ max =0 ∂ r1 r1 . . e .76 r   r  .Lucas High Voltage Cables i . ∴ ξ max = V V = 1. = 0  r1 r1 r  24 This gives the solution r1 = 1.

Lucas High Voltage Cables 25 However. . the addition of the intersheath raises the maximum applicable voltage by 33%.1 = 0. Consider the case of only overall diameter R being fixed. e . Then for minimum value of maximum stress we have ∂ ξ max = 0. Hence. and both r and r1 being variable. ∂r ∂ ξ max =0 ∂ r1 R r i . we have ξ'max = V/r.881 e r1 r1 .125 J.R. loge r1 . r1 = e. also .1 + loge + = 0 r r1 e R 1 R This gives loge = 1 giving = 1.1 + loge + = 0 r r1 r1 R 1 i . for the cable without intersheath. e .

 + r   r1  e V i .126 J.Lucas High Voltage Cables ξ max = V V V = = R 1 er e r loge + r loge e e r  1 .R.718 r . ξ max = 2. e .

and (b) applying gas pressure. the void control is effected by pressurising the oilimpregnated paper tape insulation by (a) pressurising the oil.Lucas High Voltage Cables 27 3.4Pressurised High Voltage Cables In high voltage paper insulated cables.R. . In super voltage cables.127 J. the application of pressure (about 13 atmospheres) increases the maximum allowable working stress (after applying a suitable safety factor) from about 50 kV/cm to about 150 kV/cm.

the drying and impregnating are done before sheathing.5 atmospheres or 50 p.Lucas High Voltage Cables 3. A reservoir maintains the required pressure. . The oil filled construction permits a great reduction in size of the cable. The maximum pressure of oil utilised is about 0. The cable can now operate at a maximum working stress of 150 kV/cm.128 J. Due to the pressure of oil.R. the oil must be free to flow inorder to transmit the pressure. solid type of cable.i.4.).s.35 MN/m2 (3.1 Oil-pressure cables In oil filled cables. the sheath tends to bulge out and therefore reinforcement of the sheathing is necessary. while in oil-filled cables they can be done after sheathing by circulating hot oil. In normal.

filler-space channels.129 J.24 has a hollow conductor which acts as an oil channel. and (c) three-core. Figure reinforcement helical lead paper hollow oil duct3.R.Lucas High Voltage Cables There are 3 main types of oil filled cables.24 – Single core conductor channel cable ribbon sheath insulation conductor (a) Single-core conductor channel This type of cable shown in figure 3. sheath channel. (b) single-core. These are (a) single-core. A disadvantage of this arrangement is that the oil is at high voltage . conductor channel. and is the simplest from the point of view of the cable itself.

130 J. The copper strands of which the conductor is made are laid over a helical metal ribber. so that oil can reach the insulation. (b) Single-core sheath channel .Lucas High Voltage Cables with respect to earth being at the voltage of the conductor.R.

R.131 J.Lucas High Voltage Cables Flead p grooved button-stamped oil a lead spacer ribbon sheath helical channels p sheath e r i n s u l a ti o n .

the oil channels are produced either by grooving the sheath or by arranging spacers between sheath and insulation. filler space channels .25). so that more feeding points are necessary to maintain the pressure.R. The resistance to oil flow in this type is 6 to 8 times that of type (a).132 J. An advantage is that the channels are at earth potential so that joints and installation are simpler. (c) Three-core.Lucas High Voltage Cables In this type (Figure 3.

133 J.R.Lucas High Voltage Cables F p l a e r p a f e d o r s r f i h a il n e t l s a e u t d r l h o a il ti o d n u c .

. the oil channels are located in the filler spacers.26).R.Lucas High Voltage Cables In this type (Figure 3.s 134 J. These channels are composed of perforated metal-ribbon tubing and are at earth potential.

) is used.27 -(14 Impregna Pipe Pressure Types of Gas line External Internal High Selfcontained cush pressure Pressure atmospheres) type ted pressurise gas-filled ion type d . Figure 3.135 J.4.4 MN/m2 (14 atmospheres or 200 p. Figure 3.i.Lucas High Voltage Cables 3.R.s.27 shows the different types of gas pressure cables.2 Gas-pressure cables In Gas pressure cables. a pressure of about 1.

3 External Pressure Cables Pipe line type .4.R.Lucas High Voltage Cables 3.136 J.

137 J. e i e n l m p e i t p a e l t li a n p e .Lucas High Voltage Cables S t Nitrogen lead Ftriangular at t 200 (membrane) h sheathp.s.R.i.

The pipe is filled with Nitrogen subjected to a pressure of 200 p.28.s. which is transmitted to the insulation through the diaphragm.Lucas High Voltage Cables The cable.138 J. is manufactured in the usual way and the outside is made triangular. and covered by a diaphragm lead sheath.R. shown in Figure 3. .i.

In the Self-contained type.Lucas High Voltage Cables The steel pipe is laid first. and the cable is drawn in afterwards. but otherwise the principle is the same as that of the pipe line type.R. The pressure is transmitted to the membrane through the membrane.139 J. an additional reinforced lead sheath is used. Nitrogen under pressure is then introduced into the pipe. .

R.140 J.4. (a) Gas filled cables .4 Internal Pressure Cables In the internal pressure cables. the gas is in contact with the dielectric.Lucas High Voltage Cables 3.

Lucas High Voltage Cables Alead r copper woven c Paper alloy Freinforcement n fabric u impregnation o tape sheath n b tt u b o l e n a r t r t a g a p a p e s e p a s s a g .141 J.R.

spaces are left between the convolutions so that the gas is between them. but cannot diffuse outside the insulation. The presence of Nitrogen prevents the formation of voids.Lucas High Voltage Cables In these cables (Figure 3.29). (b) Gas cushion type . The method of manufacture is such that the gas can move freely inside packets.142 J.R.

R.Lucas High Voltage Cables F s e a p l a e r d a t e g a s c u s h i o n .143 J.

Lucas High Voltage Cables (Type shown is not of much practical use but only of academic use). (c) Impregnated pressurised cable . with the result that the cable may be cut for joining without losing gas from more than a short length. a screened space is provided between the lead sheath and the dielectric. This storage is maintained by the subdivision of the screened space into a series of gas cushions by means of barriers.s 144 J. this space providing accommodation at all points along the length of the cable for the storage of inert gas under pressure. In this type.R.

R.145 J.Lucas High Voltage Cables s ( C t m l g F t ti o e a r n p t a s n p e a d c ll n e b l s h d r i s h a s w c e n e o d r a n d c p v t e o p r e h l a e n p d r f e u ) a r c b s t r c o i .

i.Lucas n In the manufacture of this type of cable (Figure 3. The cable has a mass impregnated paper dielectric and the impregnating oil is maintained under a pressure of 200 p. Special reinforcement is provided to cater for the large hoop and longitudinal stresses set up.31). The gas channel is in one of the filler spaces.e e 146 J. provision is made for longitudinal gas flow. by means of nitrogen. High Voltage Cables .R. The core is stranded and is covered with a metallised paper screen so as to obtain a completely uniform stress.s. The impregnating compounds used are suitable for the higher dielectric stresses necessary for high voltage cables.

Lucas High Voltage Cables 3. After excavation of the trench. servin Air (T back of cover A sand tile insulation conducto sheat fill3.Cross-section of Trench and buried cable depth undisturbed Figure g bedding r ground h burial .R.32 Trench )width.32.5Thermal Design of Cables Underground cables are installed in trenches of rectangular cross-section. a layer of sand is placed in it to serve as a bedding.147 J. as shown in Figure 3.

the minimum depth of burial in normal circumstances is 900 mm (36 inches). producing a temperature gradient. The excavated material is replaced in the trench and stamped to consolidate it. An underground cable carrying current will have in addition to the conductor loss.Lucas High Voltage Cables The length of cable is pulled in along the trench and covered with a further layer of sand. and for safety reasons.148 J. These produce heat which are conducted away from the cable to the surface. The minimum trench width that can be conveniently excavated is about 700 mm (27 inches). dielectric loss and losses in the sheath. . Sand free from flints and stone is employed to avoid damage to the cable serving during pulling and initial back filling. Above the cable and sand bedding are placed cover tiles to protect the cable from mechanical damage from subsequent excavation activities.R.

the heat produced by one conductor affects the other and the heat factors need to be modified. the heat produced by the circulating currents between the cables will increase whereas the eddy current losses decrease. . Thus there is an optimum spacing for cables and various alternatives may have to be evaluated before the economic arrangement is finally selected.R.149 J.Lucas High Voltage Cables When more than one single core cable is laid together (as is required for three phase systems exceeding 150 kV). When the spacing between the cables is increased.

3. so that the same cable would have different ratings depending on whether the cable is buried or not.R. the heat generated in the cable is due to (a) ohmic loss in the conductor.1 Current rating of Cables In a cable.2 Thermal Resistance . (b) the dielectric loss in the insulating medium and (c) the sheath and intersheath losses.150 J. The heat so generated is radiated to the surroundings. the factor which ultimately limits the current carrying capacity is the maximum operating temperature which may be sustained by the cable throughout its life without risk of damage or deterioration. The current that can be carried depends on the conductivity of the surrounding medium as well.5. As was discussed in an earlier section.Lucas High Voltage Cables 3.5.

R. the thermal resistance of the cable and surroundings is measured in terms of the thermal ohm. θ = maximum allowable temperature difference between core and surroundings (oC) Rθ = Effective Resistance of conductor (including effects of sheath loss) I = Current carried by conductor H = Heat produced in the core (W) .151 J.Lucas High Voltage Cables Since the flow of heat can be considered analogous to the flow of charge or current in the insulation. Consider a cable buried under the surface of the earth. Thermal Resistivity: The thermal resistivity is the temperature drop in degree centigrade produced by the flow of 1 watt between the opposite faces of a metre cube of the material. Thermal Ohm: The thermal ohm is the resistance of a path through which a temperature difference of 1 0C produces a heat flow of 1 watt.

Lucas High Voltage Cables S' S" S G = = = = Thermal resistance of dielectric Thermal resistance of cable outside dielectric S' + S" = Total thermal resistance of cable Thermal resistance of ground from cable to surroundings .R.152 J.

Lucas High Voltage Cables θ F i g u r e 3 .153 J.R. 3 3 H e a t fl .

m b u ri This gives the current rating of the cable as e d θ . the total power loss must equal to the heat produced.R.Lucas High Voltage Cables .w li From the definition. θ = ( W d + I 2 R θ ) (S + G) 154 J.W d (S + G) c I = R θ (S + G) a b l θ ∴ H= =W d + I 2 Rθ S +G i . the total temperature rise between the conductor and then surroundings is given by e θ = H (S + G) s Total power loss = dielectric loss (Wd) + ohmic loss (I2 Rθ) fr At o equilibrium. e .

155 J.Lucas High Voltage Cables In calculating the flow if heat it is useful to to remember the following analogies. Electricity Electrostatic Electromagnetic Potential Potential magnetomotive difference difference force (mmf) Electric Heat Flow H Current I Magnetic flux φ charge Q Thermal Resistance 1/Capacitance Reluctance resistance S If a method exists to study any of the above phenomena. Heat Temperature Difference θ .R. the analogous quantity can also be studied by comparison.

5.3 Thermal Resistance of single-core cable .R.156 J.Lucas High Voltage Cables 3.

157

J.R.Lucas

High Voltage Cables

q H d x1equi-temperature r2equi-potential Flines of 1 x charge lines heat flow lines flow

158

J.R.Lucas

High Voltage Cables

The analysis of this problem is similar to the analysis of the analogous electrostatic case. Figure 3.34 shows these two cases. In the heat problem, H is the amount of heat generated per unit length of cable, and in the corresponding electrostatic case, q is the electric flux flowing out per unit length. For the electrostatic case, consider a gaussian cylinder of radius x and thickness dx. D . 2 π x . l = q, so that D = q/(2 π x)

159

J.R.Lucas

High Voltage Cables

q ∴ ξ= , where ε = ε 0 ε r ε 2π x ∴ V=
r2

r1

q q 1 r . .d x = loge 1 2π ε x 2π ε r2

Considering the analogous heat flow case, Letdθ = drop in temperature across dx k = thermal resistivity
k . H . dx then d θ = 2π x kH r integratio gives θ = n loge 1 2π r2

Thermal resistance is θ/H, so that

160

J.R.Lucas

High Voltage Cables

k r ∴ thermalresistance S = loge 1 2π r2

161

J.R.Lucas

High Voltage Cables

3.5.4 Thermal resistance of three-core cables For three-core cables, the following two quip expressions are used.
 k  0.2 t    1.1 t   T + t  (i) S =  0.85 +  ln   4.15 +   + 1  Ω/ km 6π  T   T  r   where t = thicknessof belt insulation T = thicknessof conductorinsulation r = radiusof conductor  r 26 - a 6  k (ii) S = ln  3 2  6 π  3 r2 a r  where r = radiusof conductor a = radiusat which t centresof each conductorlies he r 2 = outer radiusof dielectric

162

J.R.Lucas

High Voltage Cables

3.5.5 Thermal resistance of protective coverings Since the protective covering of the cable is in the form of a cylinder, the expression is of the same form as that for a single core cable.
k r 3 - A/2 S= ln 2 π r 2 + A/2

where r3 r2 A k = = = = radius of outer covering of cable radius of lead sheath thickness of armouring thermal resistivity

163

J.R.Lucas

High Voltage Cables

3.5.6 Thermal resistance of ground around cable

164

J.R.Lucas

High Voltage Cables

charge earth Heat e harth -H Hh Fqq flow flow − i g u r e 3 . 3 5 E f f e c

o f 165 J.R.Lucas InsideEthe cable constant temperature lines would all be concentric a cylinders since the outer lead sheath is a conductor of heat. r The flow of heat would consequently be radial. t However, outside the cable, the equi-temperature lines would no longer be h concentric and the heat flow would go radially outwards from the surface of theS cable and ending up at the surface of the ground normally u (assuming that the surface of the earth is at a constant temperature). r Here again, let us analyse using the analogy of the infinitely long f carrying a charge q per unit length, placed at a distance h above conductor a the earth surface (Figure 3.35). c This has the same effect as having a charge of -q at the same distance e beneath the earth. The effect of the earth can be replaced by an equal and opposite charge on the opposite side of the surface at the same distance from the surface.
High Voltage Cables

166

J.R.Lucas

High Voltage Cables

The effects of the charge +q and -q can now be separately considered, and the results superimposed. Each charge considered separately will give rise to radial flux lines. The potential difference between the charges caused by one of the charges is given by
q 2h V= loge 2π ε r

The total potential difference caused between the charges is twice that of the individual charge. This is equal to
V= q - (- q) 2h 2q 2h = loge loge 2π ε r 2π ε r

Thus the potential difference to the neutral of each conductor is given as

Lucas High Voltage Cables q 2h V= loge 2π ε r Analogy: The temperature difference from the heat source to earth is given by kH 2h θ= loge 2π r Thus the thermal resistance of the ground is given by θ k 2h thermalresistance= = loge H 2π r .R.167 J.

R. .Lucas High Voltage Cables When applied to the practical case.168 J. This is because we have assumed the earth to be a plane of perfect conductivity (or constant temperature). Thus the modified thermal resistance G of the ground for practical application is given by k 2h G= loge 3π r A representative value of the thermal resistivity k of the soil of average moisture content is 180. it is found that the theoretical thermal resistance a found above has to be multiplied by a factor of 2/3.

Lucas High Voltage Cables 3. the heat dissipation is given by H = 2 πkr2 ( θ s . e θ s = temperatur of cable surface. For a surface in direct contact with the air. θ a = ambient temperature k = emissivityconstant w hose value varies with r2 .169 J.7 Cables exposed to air The heat dissipation of a cable exposed to the air depends on the radiation. with unrestricted ventilation. .usuallyof lead sheath.θ a )1.25 watt/cm of length where r2 = exernalradius of cable.R.5.

. floors. tanks etc. metal. The bushings have to provide electrical insulation of the conductor for the working voltage and for various over-voltages which may occur in service and also have to provide mechanical support against various mechanical forces.170 J.R.Lucas High Voltage Cables 3.6High Voltage Bushings Bushings are insulators which are used to take high voltage conductors through earthed barriers such as walls.

Lucas High Voltage Cables 3.6.1 Simple cylindrical bushing .R.171 J.

172 J.Lucas High Voltage Cables voltag e r R t L live bushi Frarthed R eradius barrier x conductor ng to s earth V .R.

Lucas High Voltage Cables The simplest form of bushing is a cylinder of insulating material around the conductor (Figure 3.173 J. so that .e. In this case.r) and axial length L to suit electrical strengths of the insulating material and surrounding media. The expressions for stress ξ and voltage to earth Vx at radius x is given by R V x ξ= . at x = r).R. with radial clearance t = (R . V x= R R x ln x ln r r V ln The maximum stress in the bushing occurs at the conductor surface (i. the voltage distribution with radius x is not linear so that the material is not equally stressed.36).

2 ξ max 2 r ξ max Thus as the voltage increases..174 J.1    max    2 V V t= + + .Lucas High Voltage Cables ξ max = V r +t r loge r so that the thickness t required for the bushing is given by   V    t = r exp  r ξ  . so much so that for very high voltages. simple cylindrical bushings of this form are not satisfactory. .R.. the dimensions required become very large..

Lucas High Voltage Cables Upto about 66 kV.R. .175 J. porcelain bushings (with or without oil) may be used.

the busing is divided into a number of capacitors by concentric cylinders of metallic foil or metallic coated paper. it is possible to obtain a nearly uniform voltage distribution with radius (Figure 3.6. .R.37).Lucas High Voltage Cables 3.2 Condenser bushing The difficulty of the dimensions increasing rapidly is overcome by the condenser bushing.176 J. In this case. By proper choice of lengths of these cylinders.

Lucas High Voltage Cables cylinders ┌┐ ││ ││ ││ ┌─ ╥─ ─┘ │ ── ── ─┴ ─╨ ── ─┴ ── ── .177 live conductor earthed flange equipotential J.R.

s. it must be equal to the ratio of the applied voltage V to the thickness t' of the insulation. V V ′= ξ max = .R. determine the radial thickness t' of the bushing. voltage of 30 kV to earth is designed to have a uniform radial voltage gradient (Figure 3.m. so that t t′ ξ max ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── As can be seen when comparing with the simple bushing.38). the thickness t' required now is much less than t. ── ── ── ── . Assuming a uniform and very small thickness of insulation between each successive foil.── ── ── ── ── ── ── ── ── ── ── ── 178 J. Example A condenser bushing for an r. The insulating material used has a maximum permissible working voltage stress of 10 kV/cm (peak).Lucas High Voltage Cables Since the stress is now made uniform.

Lucas High Voltage Cables ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ─ If the length of the bushing at the outermost radius is 10 cm. Estimate also the thickness t for the bushing without foils.── ── ── ── ── ── ── ── ── ── ── ── ── ── ── 179 J. if it is to have the same maximum radial stress.R. . determine the length at the surface of the conductor (radius 2 cm).

38 = Length of condenser bushing Figure ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── .── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── 180 J.R.Lucas High Voltage Cables 1 l 0 cm 3.

l = 62.y = a = 62. y = l therefore.2 cm In the absence of foils.4 at x = r = 2. ── ── ── ── ── ── ── ── ── ── ─ ── .Lucas High Voltage Cables Since stress is uniform.24 + 2) therefore.R. y = 10 cm.24 cm 10 ξ max ── ── ── ── ── ── ── ── ── ── ── ─ The profile of the bushing has the equation y = a/x. ── ── ── ── ── ── ── ── ── ── ── ── ── ─ V 30 2 t′ = = = 4. at x = t' + 2.4/2 = 31.── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── ── 181 J.4/x = 62. x. so that a = 10(t' + 2) = 10(4.

Lucas High Voltage Cables └─ ╨─ ─┐ │ ││ 30 2 ∴ 10 = ξ max = r +t t+2 r loge 2 loge r 2 1 ∴ loge (1+ 2 t) = 1.R.68cm 2 ││ ││ Thus in the absence of grading.5 2 = 2. there are other types of bushings. and the condenser type bushing.24 cm). e . 1 + 1 t = 8.╥─ ── ┬─ ── ── ── ── ─ 182 J. which may consist of more than one material. └┘ .342 = 14.342 ∴ t = 2 x 7.68 cm as compared to 4. it is seen that a much greater thickness of insulation is required (14.121 V i . In addition to the simple cylinder bushing.

4 Measurement of High Voltage .

Lucas 184 4.0High Voltage Measurement High voltages can be measured in a variety of ways. Lightning surges may be recorded using the Klydonograph. High Voltages are also measured by stepping down the voltage by using transformers and potential dividers. and several forms of voltmeters have been devised which can be connected directly across the test circuit. The sparkover of sphere gaps and other gaps are also used. .R. especially in the calibration of meters in high voltage measurements. Transient voltages may be recorded through potential dividers and oscilloscopes. Direct measurement of high voltages is possible up to about 200 kV.Measurement of High Voltages J.

1 Electrostatic Voltmeters One of the direct methods of measuring high voltages is by means of electro-static voltmeters. generally the attracted disc type of electrostatic voltmeter is used. Lucas 185 4.1.R.Measurement of High Voltages J. then the energy stored in the is given by . For voltages above 10 kV. When two parallel conducting plates (cross section area A and spacing x) are charged q and have a potential difference V.1 Direct Measurement of High Voltages 4.

R. so that the meter reads the square value (or can be marked to read the rms value). .2 x dx x V ∴ F=.1 Aε 2 N 2 x 2 It is thus seen that the force of attraction is proportional to the square of the potential difference applied. Lucas 186 Energy stored W = 1 2 C V 2 so that changed W = 1 V 2 d C = F d x 2 1 2 2 dC ∴ Force F = V N dx Aε dC Aε for uniformfield Capacitanc C = e so that =.Measurement of High Voltages J.

Figure 4.R.v. voltages].Measurement of High Voltages J.v.c. [The force in these electrostatic instruments can be used to measure both a. h. Lucas Electrostatic voltmeters of the attracted disc type may be connected across the high voltage circuit directly to measure up to about 200 kV.1 xkmarkings on – Abraham axis g electrod electrode electrostatic voltmeter V e 187 . sprin l.c. and d. without the use of any potential divider or other reduction method.

Owing to the difficulty of designing electrostatic voltmeters for the measurement of extra high voltages which will be free from errors due to corona effects. within the instrument.Measurement of High Voltages J. The right hand electrode forms the high voltage plate. The two large discs form adequate protection for the working parts of the instrument against external electrostatic disturbances. while the centre portion of the left hand disc is cut away and encloses a small disc which is movable and is geared to the pointer of the instrument.R. The range of the instrument can be altered by setting the right hand disc at pre-marked distances. 188 . a number of special methods have been devised for the purpose. These instruments are made to cover ranges from 3 kV to 500 kV. There are two mushroom shaped hollow metal discs. Lucas Abraham Voltmeter The Abraham voltmeter is the most commonly used electrostatic meter in high voltage testing equipment. and to the external electrostatic fields.

two metal spheres are used. and on the density of the gas. separated by a gas-gap.R.1. The breakdown strength of a gas depends on the ionisation of the gas molecules. the breakdown voltage varies with the gap spacing. for different diameters and distances. the breakdown voltage variation with gap spacing. so that the sphere gap is very useful as a measuring device.2 Sphere gaps The sphere gap method of measuring high voltage is the most reliable and is used as the standard for calibration purposes. In measuring device. have been calculated and represented in charts.Measurement of High Voltages J. and for a uniform field gap. Lucas 4. The potential difference between the spheres is raised until a spark passes between them. a high consistency could be obtained. 189 . As such. By precise experiments.

. Thus the correction for any air density change must be made.R.386   760 273+ t  273+ t  190 The spark over voltage for a given gap setting under the standard conditions (760 torr pressure and at 20oC) must be multiplied by the correction factor to obtain the actual spark-over voltage. A spark gap may be used for the determination of the peak value of a voltage wave. Lucas The breakdown strength of a gas depends on the size of the spheres. P 273+ 20  P  δ= × = 0. their distance apart and a number of other factors.Measurement of High Voltages J. and for the checking and calibrating of voltmeters and other voltage measuring devices. The density of the gas (generally air) affects the spark-over voltage for a given gap setting. The air density correction factor δ must be used.

191 .2) is almost independent of humidity of the atmosphere. but the presence of dew on the surface lowers the breakdown voltage and hence invalidates the calibrations.R. Lucas The breakdown voltage of the sphere gap (figure 4.Measurement of High Voltages J.

R. D = sphere diameter D d . Lucas 192 Fwhere d = gap spacing.Measurement of High Voltages J.

accuracy = ± 3 % 0. The limits of accuracy are dependant on the ratio of the spacing d to the sphere diameter D.R. and accuracy falls.5 D.5 mm. Lucas The breakdown voltage characteristic (figure 4. accuracy = ± 5 % . 125 mm.75 D> d > 0. as follows. the uniform field between the spheres becomes distorted. 250 mm. 500 mm.Measurement of High Voltages J. d < 0. 1 m and 2 m) 193 Figure 4.5 D.Breakdown voltage breakdow gap characteristic of sphere gaps n voltage distance When the gap distance is increased. .3) has been determined for similar pairs of spheres (diameters 62.

Lucas For accurate measurement purposes.c. d-h.R.e.v.75D are not used. characteristics voltage (mm) (kV) 194 The breakdown voltage characteristic is also dependant on the polarity of the high voltage sphere in the case of asymmetrical gaps (i. gaps where . +h.Measurement of High Voltages J. – < 4.4 FigureD & Breakdown voltage breakdown gap spacing a.v. gap distances in excess of 0.

If both electrodes are at equal high voltage of opposite polarity (i.e.R. + ½ V and . then the polarity has no effect.Measurement of High Voltages J.½ V). Lucas one electrode is at high voltage and the other at a low voltage or earth potential). D/ 2Figure 4. as in a symmetrical gap.5 DD 10 sphere gap 195 .

Since the breakdown is caused by the flow of electrons.Measurement of High Voltages J. Under alternating voltage conditions. breakdown will occur corresponding to the lower curve (i. there are two breakdown characteristics. then the positive and the negative characteristics cross over due to various space charge effects. a higher voltage is generally necessary for breakdown than when the high voltage electrode is negative.R. 196 . the a. characteristic is the same as the negative characteristic.c. But this occurs well beyond the useful operating region.e. Thus under normal conditions. Lucas In the case of the asymmetrical gap. when the gaps are very far apart. one for the positive high voltage and the other for the negative high voltage. in the negative half cycle under normal gap spacings). when the high voltage electrode is positive. However.

to obtain high accuracy. since these also affect the accuracy (figure 4.R. "The length of any diameter shall not differ from the correct value by more than 1% for spheres of diameter up to 100 cm or more than 2% for larger spheres". the two gaps should not be connected in parallel. When spark gaps are to be calibrated using a standard sphere gap.5). 197 . the minimum clearance to be maintained between the spheres and the neighbouring bodies and the diameter of shafts are also specified.Measurement of High Voltages J. Lucas In sphere gaps used in measurement. One sphere may be earthed with the other being the high voltage electrode. or both may be supplied with equal positive and negative voltages with respect to earth (symmetrical gap). There is also a tolerance specified for radius of curvature of the spheres. Peak values of voltages may be measured from 2 kV up to about 2500 kV by means of spheres.

and by the corona forming at the points before the gap actually sparks over.R. Also the effect of the variation of the humidity of the atmosphere on such gaps is much greater. a resistance is used in series with the sphere gap. 198 . of about 1 Ω/V so as to limit the current under sparkover conditions to about a maximum of 1 A. Usually. but errors are caused by the variation of the sharpness of the needle gaps. Lucas Equivalent spacing should be determined by comparing each gap in turn with a suitable indicating instrument.Measurement of High Voltages J. a series resistance must not be used since this causes a large drop across the resistance. However for impulse measurements. Needle gaps may also be used in the measurement of voltages up to about 50 kV.

since the breakdown does not occur at exactly the same value of voltage each time.R. .6). Lucas 199 breakdown 50 100 Figure Percent4. and a histogram is plotted with the peak value of impulse voltage and percentage of breakdown (figure 4. what is generally specified is the 50 % breakdown value.Measurement of High Voltages J.6 .Breakdown voltage voltage (peak) % characteristic for impulses age breakd own In measuring impulse voltages. A number of same value impulses is applied and a record is kept of the number of times breakdown occurs.

2 Transformer and potential divider methods of measurement 4. Vt sup ent transformer ratio devi limit ply ing ce method resi stor 1 Ω /V 200 .1 Transformer ratio method Curr Tes a. Figure 4.c.Measurement of High Voltages J. Lucas 4.R.2.

R. In this method (figure 4. This method measures the rms voltage. Curr Test a. currents such as due to corona may cause considerable error in the measured voltage. Figure 4.7). The actual voltage across the load is not measured.with additional V devi sup ent potential winding limiti ply ce ng resis tor 1 Ω /V 201 . Since the current taken by the device under test is usually very small.8 . Lucas The use of primary voltage to estimate secondary voltage is a fairly rough method of measurement.c. the voltage on the low voltage side of the high-tension transformer is measured. In order to determine the peak value it is necessary to determine the wave form of the secondary voltage. but is satisfactory enough for most ac tests.Measurement of High Voltages J.

Measurement of High Voltages J. Lucas Some high voltage transformers (figure 4. however this is an expensive arrangement (figure 4. It may also be possible to have a potential transformer connected across the test device and the voltage measured. This method cannot be used with the cascade arrangement of the transformers.R.9).8) carry a separate voltmetercoil having a number of turns which is a definite fraction of the secondary turns. 202 .

Lucas 203 Curre Range Test a.with potential V– suppl devic 0 nt transformer 100 V limiti y e ng resist or 1 Ω /V .v. h.c.R.9 .Measurement of High Voltages J. potential Figure 4.

Lucas Even this method may not be very satisfactory under very high voltage conditions and the series resistance method of measurement may be used.R.Measurement of High Voltages J. 204 .

Method is applicable for both ac and dc. which is earthed for shielding purposes.high Safety 50 μA series v. and to prevent leakage current. gap resistor microammeter resista move ment nce (20 kΩ/V) 205 In the series resistance method a high series resistance (specially designed to withstand high voltage) and resistance of 20 kΩ/V. Lucas Series resistance method of measurement μFigure 4.R. A number of resistances would be necessary in series.Measurement of High Voltages J. . we would have to have the whole system in a insulated container. is used with microammeter (having a 50 μA movement).Series A h. .

Lucas As a safety measure. With a stable supply (accuracy 0.Measurement of High Voltages J. 206 .R. a safety gap or neon lamp is connected across the micro-ammeter.10%) final accuracy of 1% is achieved.

R.v.11 .Resistive V resistive devi sup ent transforme potential divider method limiti ply cedivider r ng(≈ 20 resis kΩ/V) tor 1 Ω /V . Lucas 207 Test Curr a.c. Figure 4. h.Measurement of High Voltages J.

there would be the effect of the distributed capacitances as well. The capacitive effects can be reduced by providing a suitable screen. or by balancing the capacitance.Measurement of High Voltages J.R. 208 . Lucas When the above method is used for alternating voltages.

209 Under alternating conditions there would be distributed capacitances. and a definite fraction of the total voltage is measured by means of a low voltage voltmeter.2. . Lucas 4.Measurement of High Voltages J. a high resistance potential divider is connected across the high-voltage winding.2 Resistive potential divider method In this method.R.

Screening of resistive dividers (b) (a) One method of eliminating this would be to have a distributed screen of many sections and using an auxiliary potential divider to give fixed potential to the screens.R. Lucas 210 Figure 4.Measurement of High Voltages J. . .

so that if at any point the capacitive current from conical screen to the point is equal to that from the point to the earth. then the capacitances would have no net effect. It also possible to have a metal conical screen (Figure 4. There would be capacitances to the conical screen as well as capacitances to earth. The design has to be done by trial and error. Lucas The currents flowing in the capacitances would be opposite in directions at each half of the screen so that there would be no net capacitive current (Figure 4.Measurement of High Voltages J.R.12 (b)). 211 .12 (a)).

3 Capacitive potential divider method . Lucas 212 4.2.Measurement of High Voltages J.R.

h. 4.Capacitive electrost stray igure 2 suppl devic nt1 transformer divider method capacitor capaci limiti y e atic voltmete tors ng resistr or 1 Ω /V V potential . .v. Lucas 213 Curre Test a.c. FC h.Measurement of High Voltages J.R.v.

and we could lump them up with the capacitances of the arms. In this two capacitances C1 and C2 are used in series. If the system is kept at a fixed position. the capacitance to the screen would be a constant. we could use a capacitive potential divider. Lucas For alternating work. instead of using a resistive potential divider.V = .R. Neglecting the capacitance of the voltmeter (or lumping the electrostatic voltmeter capacitance with C2) the effective capacitance of C1 and C2 in series is C1C2/(C1+C2). Or if screens are used.V (C1 + C2) / C1 C2 C1 + C2 214 . and since the charge is the same.Measurement of High Voltages J. 1 / C2 C1 Voltageacross C2 = . the electrostatic voltmeter being connected across the lower capacitor. we can make corrections for the fixed stray capacitances.

s. This method also measures the r.s.since it is the peak value of the applied voltage which produces the actual breakdown stress in the material under test.m. For this reason. value. 215 . air capacitances are always used for this purpose. standard capacitor must be accurately known. voltage. It is often more satisfactory however.v.s. Lucas The capacitance of h.Measurement of High Voltages J. If the shape of the voltage waveform is known.m.m. It is sometimes more useful to have a measure of the peak value of the alternating voltage rather than the r. to use some method of voltage measurement which gives the peak value of the voltage directly. and the capacitance must be free from dielectric losses. value.R. the peak voltage may be obtained from the r.

14). Lucas 4.Measurement of High Voltages J. and also to cause a small delay between the arrival of the trigger pulse and the waveform (Figure 4. 216 .2.4 Matching of Potential dividers When waveforms are observed on the oscilloscope. through a potential divider. a cable is necessary to connect the test waveform to the oscilloscope.R.

R.Measurement of High Voltages J.Observation of impulse waveform through Z2 E1 cable sweep Tim 0 4 3 plate impul under potential divider C2 e s base se test . Lucas 217 XYDevice input Figure single pulse trigger 4. .

e: the ratio of the potential divider does not vary with time or frequency). The requirement of the potential divider used are that it reduces the applied voltage without producing any distortion (i. Velocity of the wave in cable = 1 ≈ LC 1 218 µ0 ε = 3 × 108 εr The oscilloscope can display a maximum of about 50 V to 100 V and thus the impulse voltage must be reduced by a suitable potential divider. (i) Resistive . In a lossless cable . then it may be represented by purely inductances and capacitors. Zo is purely resistive.R. so that the surge impedance of cable or delay network = Zo = [z/y]1/2.Measurement of High Voltages J. The potential divider can be of two types. Lucas If the delay cable is lossless .

R. but a mixture of both. Lucas and (ii)Capacitive.R.O and the voltage. The ratio of the divider is determined by the sensitivity of the C. neither case is obtained in the pure form.Measurement of High Voltages J. In practice . the lower arm of the divider has its resistance fixed by the surge impedance of the cable used (for matching ) and by the wave–tail requirements of the impulse generator circuit (if any impulse generator is used). The capacitive effect in the resistive divider is much more than the resistive effect on the capacitive divider. 219 . but has several advantages. In the case of the resistive divider . It can be used as part of the wavefront forming circuit. The capacitive divider is generally bulkier than the resistive divider .

but when the line becomes charged. When the initial part of the surge enters the cable.O adds to the capacitance of the cable connecting the divider to the oscilloscope adds to the capacitance of the lower arm. it acts as a transmission line and presents its surge impedance to the surge. When using potential dividers.R.Measurement of High Voltages J. Lucas The self capacitance of the cable connecting the device to the C. it is necessary to suitably terminate the cable at the two ends so as to have perfect matching of the cables at two ends. 220 .R. it behaves as a capacitor.

Measurement of High Voltages J. (1) Matching at potential divider end only: 221 .R. Lucas Resistive potential dividers There are three ways in which they may be matched to delay cables.

R.Measurement of High Voltages J. Lucas 222 XY-Ti Device input single trigger 4.15 – Matching of resistive Z E cable R3 Figure plate impu under sweep at sending end only pulse Cme divider 0 3 4 2 1 2 slse test bas e .

Lucas 223 R2 Figure 3 4. the equivalent impedance of the section before the cable must be ZO Impedance = R3 + R1//R2 = R3 + R1.R2/(R1 + R2) = ZO for perfect matching at s.R. For perfect matching at the sending end. and perfect reflection at the receiving end.16 1 In this arrangement.Measurement of High Voltages J. the receiving end (ie: end connected across CRO Y-plates) is kept on open circuit. We try to obtain perfect matching at the sending end so that there is no reflection.e .

Measurement of High Voltages J.  O = R2 + R3 Z 224 At the junction of the divider E2.R. Lucas If R1 >> R2 as usually is. then we have R1//R2 ≈ R2 . the equivalent impedance to earth Z1 is given by .

. E1 = . so that the voltage transmitted to the CRO is 2 E3.Measurement of High Voltages J.R. Z0 R 2 (R 3 + Z0) . E1 2 Z0 ( Z1 + R1) Z1 + R1 Z0 .  Z 1 = R2 / / ( R3 + Z 0 ) = Z 0 = R 2 + R3 ( R2 + R3 + Z 0 ) 2 Z0 Z1 R 2 (R 3 + Z0) ∴ voltageat junction E 2 = . Lucas 225 R 2 ( R3 + Z 0 ) = R2 ( R3 + Z 0 ) . . = R2 so that E3 = = . E1 E2 E1 (R 3 + Z0) 2 Z0 ( Z1 + R1) 2 ( Z1 + R1) R 3 + Z0 This voltage waveform E3 travels towards the receiving end and is reflected at the open end without change of sign.

where R 2 (R 3 + Z0) ∴ Er = E4 = E1 Z1 = 2 Z0 Z 1 + R1 OR If the lower arm itself is balanced.Measurement of High Voltages J. R2=Z0 . i. then R3 = 0 and the voltage transmitted to the oscilloscope is given by Er = Z0 .e.R. E1 1 R1 + 2 Z 0 . . Lucas 226 R 2 .

R. Lucas (2) Matching the cable at the oscilloscope end only: 227 .Measurement of High Voltages J.

. Lucas 228 XY-Ti Device input single trigger Z E cable 4.Matching of resistive R Figure plate impu under sweep pulse at receiving end only Cme divider 0 4 3 k 2 test 1 lse s bas e .Measurement of High Voltages J.R.

. and since there is proper matching at the receiving end. Equivalentimpedanceat E 2 = Z1 = R 2 / / Z0 = R 2 Z0 R 2 + Z0 Z1 .O than when only the divider end is matched. E1 E1 ( R1 + R 2 ) Z 0 + R1 R 2 Z 1 + R1 R 2 Z 0 + + ( R 2 + Z 0 ) R1 For given values of R1. = R2 Z 0 R2 Z 0 .R.R. the cable is matched only at the receiving end so that there will no reflection at this end. E1 = . ∴ E2 = E1 Z1 + R1 229 Since the cable is properly matched at the receiving end. ∴ E4 = E2 = Z1 . R2 and E1 this arrangement gives smaller voltages at the C. it is transmitted with out any reflection. R4 = Z0 The voltage wave E2 travels along the cable.Measurement of High Voltages J. Lucas In this arrangement.

Measurement of High Voltages J.R. 230 . (ie. Lucas If the point E2 is not connected to earth through the resistance R2. if R2 =  then we have ).

Matching of resistive divider at receiving Z3 E2 cable RTim Figure 0 plate impul under sweep pulse Ck e only (with R = ∞) end 4 test 1 se 2 s bas e .R. Lucas 231 XYinput Device single trigger 4.Measurement of High Voltages J. .

Measurement of High Voltages J. Lucas For this case the voltage at the oscilloscope is given by 232 .R.

R. Lucas (3) Matching at both ends E4 = R2 Z 0 . E1 Lt∞ ( R1 + R2 ) Z 0 + R1 R2 Z 0 + R1 R2 → 233 of the cable : . E1 = Z 0 .Measurement of High Voltages J.

Lucas 234 XY-Ti 4. .Matching of resistive Device input Figure single trigger Z E cable R plate at both ends impu under divider sweep pulse Cme 0 4 3 k 2 test 1 lse s bas e .Measurement of High Voltages J.R.

Also. R2 + R3 = Z0.R. Lucas In this case. at E2. for perfect matching at receiving end. the cable is matched at both ends . This arrangement is used when it is necessary to reduce to a minimum the irregularities produced in the delay cable circuit. As before. With this termination there is no reflection at either end.Measurement of High Voltages J. the equivalent impedance Z1 to earth is given by 235 . R4 = Z0 and for perfect matching at sending end.

E1 = . E1 2 ( R1 + Z 1 ) .R.Measurement of High Voltages J. E1 2 Z0 ( Z1 + R1) Z1 + R1 Z0 . E1 E2 E1 (R 3 + Z0) 2 Z0 ( Z1 + R1) 2 ( Z1 + R1) R 3 + Z0 Due to perfect matching at the receiving end. Lucas 236 ( + ) ( + ) = R 2 / / ( R3 + Z 0 ) = R 2 R3 Z 0 = R 2 R3 Z 0 . = R2 so that E3 = = . this is transmitted without any reflections. Z0 R 2 ( R 3 + Z0 ) . ∴ E4 = E3 = R2 .  Z 0 = R 2 + R3 Z1 ( R2 + R3 + Z 0 ) 2 Z0 Z1 R 2 ( R 3 + Z0 ) ∴ voltageat junction E 2 = . .

R. Lucas 237 c s Figure 4. .Measurement of High Voltages J.Stray capacitances and chopped wave h tr o a p y p c e a d p w a a c v it e a n c e s t .

then the errors due to the capacitances are large (especially in waves such as chopped waves). By having a distributed capacitance along the resistance which are larger than the stray capacitances. Capacitive potential dividers The effect of stray capacitance may be made constant. An easier way to compensate for the stray capacitances is by having capacitive potential divider instead of the resistive divider. by shielding the potential. When the rate of change of voltage is high. in a capacitive divider. the effect of the stray capacitance may be eliminated. and hence make an allowance for it. 238 . divider. Lucas The stray capacitances present between the turns of the resistances would make the current distribution along the resistance non–uniform.Measurement of High Voltages J.R.

239 .Measurement of High Voltages J. Lucas The disadvantage of the capacitive potential divider is that proper termination cannot be done. There are two methods used to couple capacitive dividers to delay cables.R.

Measurement of High Voltages J. Lucas (1) Simple capacitor connection 240 .R.

= 4.Matching of capacitive divider input Figure Z R E1 cab C3 er plate capacitor connection) impu (simple leC Zlse 0 4 k 2 1 s0 2 . Lucas 241 trigg Y.Measurement of High Voltages J.R. .

we attempt to prevent reflections at sending end. C1 = E1 . E3 travelling towards it would be reflected and hence the voltage transmitted to the CRO would be doubled. Z 0 = 2 E 2 . The sending end is terminated with a resistance R1 = Zo in series with the cable. 242 . E2 C1 + C 2 1 = E 2 . Initially the cable capacitance would not have charged up.R.Measurement of High Voltages J. Lucas In the simple capacitor connection. Initially. and only C 1 and C2 would be present.  for matching R1 = Z 0 E3 Z 0 + R1 Due to perfect reflection at the receiving end.

.R. Lucas 243 i . As time goes on. voltage at Y–plates would be given by Figure 4. 2 C1 C1 + C2 This gives the amplitude of the voltage wave as it reaches the Y–plates. the cable capacitance charges up and behaves as a capacitance in parallel with the lower arm. E r = 2 E3 = 2 x 1 E 2 = E 2 = E1 . after infinite time. e .Waveforms for simple ti E → capacitor connection m delay 4 e ← .Measurement of High Voltages J. Therefore.

Measurement of High Voltages J. end of the delay cable and connecting it in series with a resistance equal to the cable surge impedance Z0 (resistive if cable is lossless). Lucas 244 C1 C1 + C 2 + C k Thus the ratio of the input voltage to the output voltage of the capacitive divider varies with time and we get a distorted output waveform displayed on the oscilloscope. Thus the capacitive potential divider introduces distortion. The difference between the initial and final ratios will be appreciable unless C2 is at least 10 times that of the cable capacitance Ck. (2) Split capacitor connection E r = E1 . This error can be reduced by transferring part of the low voltage capacitor to the C.R.R. in which case the error would be about 10%.O. .

.R.Measurement of High Voltages J. Lucas 245 Figure Input x cable 4.Matching of capacitive divider (simple capacitor yo Z4 E3 R2 C13 2 1 connection) impulse plates Ck .

Further to ensure that the long term ratio remains the same as the initial ratio.Measurement of High Voltages J. the voltage wave is transmitted without any reflection.  R1 = Z0 for matching Due to perfect matching at the receiving end. in addition to matching the cable at the sending end (R1 = Z0). it is also matched at the oscilloscope end (R2 = Z0). Initially E 2 = E1 . Initially the capacitances Ck and C3 would not have charged. C1 C1 + C2 1 2 also E3 = E 2 . Z0 = R1 + Z0 E 2 .R. the lower end capacitor is split into C2 and C3. Lucas 246 In this connection. and only the capacitances C1 and C2 would be effective in the voltage ratio. .

. Lucas Therefore the observed voltage is given by ∴ Er = E3 = 1 2 247 E1 .R. C1 E2 = 2 C1 + C 2 After infinite time. since C3 would no longer be conducting. the capacitances Ck and C3 would have completely charged up. and the receiving end in effect would be on open circuit.Measurement of High Voltages J.

Measurement of High Voltages J. .Waveforms for split capacitor plate impu leC connection 0 k 3 2 1 lse s 2 Since all the capacitors C2.R. . Lucas 248 Yinput Z R4 E3 cab C1 Figure 4. C3 and Ck are in parallel.

C1 + C2 = C3 + Ck If this condition is satisfied. C1 C1 + C 2 + C 3 + C k If the initial and the final values of the ratio are made equal. . Lucas 249 E 2 = E1 . C1 E1 . then the distortion is reduced to a great degree.Measurement of High Voltages J. E1 2 C1 + C 2 C1 + C 2 + C 3 + C k i .25. e .R. C1 = . then the distortion is low and near faithful reproduction can be expected as shown in figure 4.

.R. Lucas 250 tim E delay → e ← 4 4.Measurement of High Voltages J.3Measurement of Surges 4.1 Klydonograph Lightning is probably the most spectacular of the high voltage phenomena.3.

as it is not possible to create lightning or to obtain a lightning strike when and where we please. photograp dielec li h.R. Also very little is known of its effects and the voltages of the surges that appear in the transmission lines due to it.Measurement of High Voltages J. The frequency of occurrence of surge voltages and the magnitude of the surge it produces on the transmission lines could be studied using Litchenberg patterns obtained by using a Klydonograph. The phenomena of the lightning could be studied to a certain extent by the surges it produces on the transmission lines.v. metal 4. Lucas Very little is known about lightning. Figure hic tric ne film electrod electrode Klydonograph e 251 .

on the surface of which is placed a photographic film. The high voltage is applied to the pointed electrode and the other electrode is generally earthed. When an impulse voltage is applied to the high voltage electrode. The insulator material separates a plane electrode on one side.R. Lucas The Klydonograph (Figure 4. the resultant photograph shows the growth of filamentary streamers which develop outwards from the electrode. and a pointed electrode which is just in contact with the photographic film. The apparatus is enclosed in a blackened box so as not to expose the photographic film. The photographic film can be made to rotate continuously by a clockwork mechanism.26) has a dielectric sheet. 252 .Measurement of High Voltages J.

R.Measurement of High Voltages J. then the film would become exposed and no patterns would be obtained. When a positive high voltage is applied to the upper electrode. Lucas This imprint on the photographic plate is not due to normal photographic action. and occurs even through there is no visible discharge between the electrodes. If flashover of the insulator or a visible discharge occurs. clearly defined steamers which lie almost within a definite circle is obtained. These patterns obtained on the photographic film are known as Litchenberg patterns. 253 .

the radius of the pattern obtained (Figure 4. Thus by calibrating the Klydonograph using a high voltage oscilloscope and known surge voltages.R. to v.27a) is a definite function of the voltage applied. radius of Klydonogr conductor positive negativ (mm) age Klydonograph pattern aph patterns e (kV) 254 patterns For a given apparatus with a fixed thickness of dielectric.Measurement of High Voltages J. For both types of surges. then the observed pattern is blurred and the radius of the pattern is much smaller. Lucas If the voltage applied is negative. . volt Figure 4. forfor h. it is possible to use this apparatus to record surges that occur. the radius of the pattern obtained increases with increase in voltage.

255 . Thus usually it is preferable to use the positive pattern for the measurement of high voltage surges. Lucas If the positive voltage applied is increased beyond a certain value.27b) For a fixed apparatus. an insulator string potential divider is used.such as occuring in transmission lines. Thus to measure voltages beyond this value.R. for a positive high voltage applied as the top electrode. The maximum voltage that can be measured using a Klydonograph is dependant on the thickness of the dielectric material.Measurement of High Voltages J. branching may occur along the branches coming out from the electrode. the variation of the applied voltage with radius of the pattern obtained is quite definite and the radius is quite large. (Figure 4. the characteristics is much more variable and the radius is much smaller. In the case of the negative high voltages.

we make a modification to the apparatus.Measurement of High Voltages J. positi small large Figure positive positi ve voltage Litchenberg patterns voltag ve voltag e e 256 Since the surges due to lightning may be either positive or negative. Lucas The applied voltage versus radius of pattern characteristics of the Litchenberg pattern is shown in figure 4. negati 4. and since it is preferable to observe the positive pattern in either case. .R.28.

there are two such instruments. Lucas 257 Negative Pattern Positive Pattern In the modification shown in figure 4. and a negative pattern in (b). . then a positive pattern would be recorded in (a).R. Thus in the modification.29.Measurement of High Voltages J. if a positive surge comes. with the electrode connections made in opposite directions. of which the pattern on (b) can be used for the measurement of the positive surge.

) e electrode 258 for measurement of both polarities .Measurement of High Voltages J. Lucas In the case of a negative surge. the opposite would happen. and the pattern (a) (b lin h.v.R. earthelectrode Klydonograph Figure 4.

it is possible in some elaborate apparatus to record the date and time occurrence of the surge as well. Since the photographic film is continuously moving.Measurement of High Voltages J.R. 259 . Lucas Thus the magnitude of the surge as well as the polarity could be determined from the Litchenberg patterns on (a) and (b).

Capacitor a.R.Measurement of High Voltages J.c.4General measurements 4. .4. dio C tim devi ent charging resista method sup de e ce limit nce ply ing R resis tor 1 Ω /V . Lucas 260 4.1 Peak reading voltmeters (i) Capacitor charging method Test Curr Figure high 4.

. (ii) Using neon lamp neo Figure suppl 4.R. Lucas 261 In the positive half cycle. l.Measurement of High Voltages J. capacitor charges up to peak value. capacit capacito n using neon lamp y or r (fixed) lam (variabl p e) A neon tube (if the voltage at which the lamp strikes is known) can be used with a capacitive potential divider to obtain the peak value of an applied voltage waveform. and when the voltage falls it discharges (very slightly) through the milliammeter.v. .Measurement h.v. so that the voltage across the capacitor is very nearly a constant at the peak value and the current is thus proportional to the peak value. (Time constant RC must be very high in comparison to period of applied voltage).

Since the extinction voltage is more constant than the striking voltage. the supply voltage can be calculated.c. An accuracy of ± ½ % could be obtained with the striking voltage and an accuracy of ± ¼ % could be obtained with extinction voltage. .Measurement of High Voltages J.v.Rectifier-capacitor current b de capacito V meter method r 262 . supply d.R. From the ratio of the capacitances. (iii) Rectifier-Capacitor current method Figure + i– dio h. the extinction voltage could be used as the standard. Lucas The low voltage variable capacitor is varied until the neon lamp strikes. 4.

R.Measurement of High Voltages J. 263 . A high voltage capacitor is connected to the hv supply with a rectifier ammeter in the earth connection. The diode used in series with the milliammeter should have a low forward resistance and a high reverse resistance a ratio of 1:105 is desirable. The indicated value will correspond to the peak value of the positive or negative half cycle. Silicon diodes provide an ideal rectifier for the purpose. Lucas The best known and the most usual method of measuring the peak value is the rectified capacitor current method.

Measurement of High Voltages J.R. Lucas 264 Figure 4. .Waveforms for (b) ( t V v i a) A non-sinusoidal B m Sinusoidal τ peak measurement Waveform a x .

. d V dt A A τ . giving V max = = Iav 2C 2Cf + d.R. Iav ∴ Iav . d t = ∫ C .c. τ = C . so that ∫ i . 2 V max . Supp Figure 4.Full wave ly − circuit . Lucas 265 B B dv i= C .Measurement of High Voltages J.

Instead of using a half wave rectifying unit as in figure 4.34. In such a case the meter reading would no longer corresponds to the actual maxima (Figure 4. independent of the waveform. and hence would correspond to the maximum value of voltage.R. Lucas Since a d.33(b)). but an addition of successive peak-to-peaks. In this case.32. the reading of the meter would effectively be double giving the result V max = I av 4C f 266 . except in the case when there is more than one maxima and minima per cycle. we could also used a full wave rectifying unit as shown in figure 4.Measurement of High Voltages J.c meter is used it would read Iav.

we can obtain the peak value of the voltage independent of the wave form.Measurement of High Voltages J. Lucas Thus using either half wave or full wave rectifying units.R. 267 . if the capacitance and frequency are known from the reading of the d.c meter.

(a) if it is stationary. Lucas 4.2 Oscilloscope for measurement of fast transients 268 High voltage oscilloscopes are used for the study of fast transient phenomena. particularly in the work on high-voltage and on spark breakdown in small gaps. . These have a high sweep speed. Since the speed is high.Measurement of High Voltages J. In these the beam should not come on till the transient comes in because. the spot of high intensity would fog the photograph before the transient comes on. the intensity is lowered and hence a higher intensity is required.R. and (b) if it is moving. by being triggered by the transient. Thus the beam should be brought on just before the transient comes on. the beam may have swept before the transient comes.4.

5Measurements of capacitance and loss tangent . The delay cable ensures that the transient appears slightly after the beam comes on. and triggered by the signal. as transients are not repetitive.the anode – cathode voltage should be high (50 – 100 kV).3 μs. so that the whole transient is clearly seen. The sweep generator should produce a single sweep (not repetitive). Lucas The transient should come on the Y-plate only shortly after the beam. For this a delay cable is used. 269 4.R.Measurement of High Voltages J. For high writing speed. Such a scope can have a maximum of 50 V to 100 V applied across Y plate so that we would have to use a potential divider. The delay cable causes the signal to appear at the Y–plates a fraction of a micro second after the sweep generator is triggered (100m length of cable may cause a delay of about 0.

Safety Figure 4.Measurement of High Voltages J.5. Lucas 4.R.1 High Voltage Schering Bridge High Voltage Schering Bridge is the method most widely used for measuring capacitance and loss tangent (or power factor) in dielectrics. S Q High C High DSafety Pa.c. Voltage gap Voltage gap 3 2 1 sup Capacitor 100 V 100 V Standard ply Capacitor 270 - High Voltage .

R. 271 . For reasons of safety. and a parallel combination of a standard low resistance S and a variable capacitance C3. The other components are at low voltage and are not allowed to have voltages greater than about 100 V applied across them by means of safety gaps connected across them (The safety gaps are either gas discharge gaps or paper gaps). The other 3 arms are a standard high voltage capacitor C2 (generally a loss free air capacitor of value 100 to 500 pF). Lucas One arm is the high voltage test capacitor (represented by a series combination of capacitance C1 and resistance P for ease of analysis). a variable low resistance Q. The high voltage supply for the bridge is obtained through a high voltage transformer.Measurement of High Voltages J. only the high voltage test capacitor and the high voltage standard capacitor will be at high voltage.

usually of the moving magnet type (If the moving coil type is used.Measurement of High Voltages J. it has to be tuned).R. Lucas The impedance of these arms must thus necessarily be of values much less than that of the high voltage capacitors. the detector used is a vibration galvanometer. The arms Q and C3 are varied to obtain balance. Figure I V ϕ δ 4. 272 . For measurements at power frequencies.

.Measurement of High Voltages J. the detector would also be near earth potential.φ = θ θ ≈ tan θ = ω P C1 = ω C3 S The low voltage end of the bridge is usually earthed. loss angle = π/2 . Lucas It can be shown that this bridge is frequency independent. Thus all the variable arms and the detector can be safely handled by the operator.R. and that at balance 273 Q C2 = . S C1 P C also = 3 Q C2 power factor angle = φ. and since the voltages across Q and S are limited to about 100 V.

so that the relative sensitivity will be small. However. 274 .Measurement of High Voltages J.R. this is not a practical disadvantage and a reasonable variation can be obtained across the detector. since the applied voltage is high. Lucas It should be noted that the bridge is an unequal arm bridge.

R. .Bridge with guarded standard S Q C D P a.c. Lucas 275 b aFigure 4.Measurement of High Voltages J. ′supply ′ 3 2 1 .

The following procedure is used to have the guard electrode at the same potential as the main electrode.R.the normal Schering Bridge. The bridge is adjusted for balance with the switch in position (a) . or kept at the same potential as the main electrode without a direct connection as shown in figure 4.Measurement of High Voltages J. Lucas Since the value of the standard capacitor must be accurately known. there should be no distortion of the field in it. Then with the switch in position (b) the bridge is again balanced using only S' and C3'. Thus a high voltage guard is provided in it design. This guard is earthed directly (which causes a small error). This ensures that finally a and b are at the same potential (same potential as the other end of the detector). 276 .37.

R. 277 .Measurement of High Voltages J. This connection can be used for capacitances up to 2000 pF. Lucas Successive balance is carried out in positions a and b alternately until final balance is obtained.

Lucas 278 a p r S Q C D P.Schering Bridge for high capacitances supply 3 2 1 .R.c.Measurement of High Voltages J. . Figure 4.

The expression at balance is obtained by converting the mesh consisting of r.R. it can be shown that 279 .Measurement of High Voltages J. At balance. p and Q into star form. In this case we have a high current fixed value resistor shunted by a low current variable high resistance which acts similar to a potential divider. thus obtaining the normal schering bridge arrangement. Lucas When it is required to obtain higher value unknown capacitances (such as in the case of a very long cable). the circuit is modified in the following manner so that high current variable resistance standards would not be required.

R. Then all the equipment. The earthing of the test capacitor near the detector end instead of the source end would bring the instruments near earth potential. The operator can either operate the instruments using long insulated rods. C1 = C 2 Q p  r so that tanθ = ω C3 S 1 . or get into a Faraday cage (A cage which is raised to the same potential as the high voltage electrode so that there is no difference in potential).Measurement of High Voltages J. and hence the operator would have to be at a high voltage to earth. Lucas 280 S Q+r+ p . then we would have to earth the end of the source near the test capacitor. but is not .C 2   C3 p  In the case of a cable already buried or earthed.

281 .R.Measurement of High Voltages J. Lucas used due to the introduction of stray capacitances by this means which would cause measuring errors.

. Lucas 4.5.loss angle 282 .capacitance of sample X Y δ 1C .Measurement of High Voltages J.Dielectric loss measurement using oscilloscope C1 Vs2 IS S .loss free capacitor (C >> C ) s s δ .R.2 Dielectric loss measurement using Oscilloscope Figure 4.

Lucas In an oscilloscope. When the two voltages are in phase. the area increases and reaches a maximum when the phase angle difference is 90o. A potential difference proportional to the applied voltage is applied to one pair of plates and a potential difference proportional to the integral of the current through the dielectric is applied to the other pair.Measurement of High Voltages J.R. This property is made use of in dielectric loss measurements. Voltages across capacitor and across sample applied across the 2 plates. the figure will be a straight line with an enclosed area of zero. Since the loss is to be measured in a dielectric sample. As the phase angle increases. if two alternating voltages of the same frequency are applied to the x and y plates. 283 . the resulting figure will be an ellipse. a lossless large capacitor is connected in series with the sample.

δ) = a . 284 . b are constants. V1m sin (ω t . Lucas The area of the ellipse thus formed is proportional to the power loss in the dielectric. The use of the standard capacitor C ensures that the voltage across it is 90o out of phase with the current.δ) and x = b .δ). the figure traced out on the oscilloscope would be a straight line. (Ism/ω C) sin (ω t . Power loss in Cs = V2 Is sin δ on the oscilloscope y-deflection ∝ v1 = V1m sin (ω t . and x-deflection ∝ v2 = V2m sin ω t. taken as reference.e. y = a . If the power loss in the dielectric is zero. Hence the angle on which the area of the ellipse depends is not the power factor angle but the loss angle.R. i.Measurement of High Voltages J. V2m sin ω t where a.

dt ∫ ωC 0 a .R.δ ) .Measurement of High Voltages J. sin( ω t . I sm . cos ω t . ω . I s V 2 sin δ C ω T It is thus seen that the area of the ellipse is proportional to the power loss. dx = a . . b .V 2m . Lucas The area of the ellipse traced out on the oscilloscope screen is given by 285 A = ∫ y .b 2 π = . .

3 Detection of internal discharges It can be done by 286 Detection of internal discharges can be carried out by various methods. Lucas 4.the audible clicks given out by the discharges may be detected by using a microphone.in transparent insulation the sparks can be detected by either direct observation or by using a photo-electric cell. these will be detailed out in the following . (a) visual methods .5. and (c) electrical methods sections.R.Measurement of High Voltages J. (b) audible methods . an ultrasonic detector or other transducer.

coro bloc Figure arge men voltag ing wave king na x b q e capac gener dete capa itor citor ctor ator Discharge 287 The discharge detector shown in the figure 4. is basically a wide band amplifier with a gain of 106 and a bandwidth of 10 kHz to 150 kHz. Lucas Electrical Methods of discharge detection (a) Using a corona detector e qstep C disch speci test coupl 4.R.40. it may be assumed that the high voltage supply circuit provides almost infinite impedance. When the corona detector shows no discharge across it. and that the step wave generator has a negligible internal impedance.Measurement of High Voltages J. Thus the discharge flow path is as shown on the diagram. The lossy dielectric sample may be represented by a capacitance Cx in parallel with its discharge qx. For the charge flow. the voltage drop caused by the coupling capacitor Cq must equal the voltage produced by .

Lucas the step wave generator. eq q Cq = 0. q  +  =   Cx  Cb C x  + (q . q = eq Cq 288 also Cb  1 qx 1  i. q i . and the voltage across the blocking capacitor Cb and by the specimen must must be zero.R. the total discharge through the specimen in the direction of q.q x ) / Cx = 0 . e . Since the specimen has its own discharge in the opposite direction to q.Measurement of High Voltages J. Since no drop occurs across the detector at balance. must be q qx.e.

Lucas 289  Cx  q x = eq C q  1 +    Cb   Substituting for q.R.Measurement of High Voltages J. . we have The energy dissipated in the void is given by w = ½ qx V 0 where V0 = peak voltage across specimen at inception voltage (b) Using the oscilloscope with filtration and amplification Internal discharges occurring within dielectric samples can be observed by measuring the electrical pulses in the circuit where such discharges occur.

Measurement of High Voltages J.R. transformer impedance h. Lucas 290 CRO h. Z Circuit dielectric high loss-free Figure coupling pass sample filter capacitor k for discharge detection .v.f. amplifier choke 4.

291 .R.Measurement of High Voltages J. Lucas The apparatus used in the observation (namely the coupling capacitor and the impedance) should be discharge tree. so that all the discharges caused is due to the sample.

. Lucas 292 h.R.Output waveforms pulses superposed on sine wave (power freq) .f.Measurement of High Voltages J. low amplitude high frequency Figure 4.

R. 293 . The discharge pulses caused in the sample are of high frequency. In the absence of this low impedance path. Lucas However. The coupling capacitor k is provided so that the high frequency components would be provided with a low impedance path.Measurement of High Voltages J. (c) Using oscilloscope with elliptical time base In many instances. so that we bypass the low frequency and amplify the high frequency in the measurement circuit. and matching units are employed which permit the use of about 30 m of co-axial lead between detector and the source of discharge. the path is highly inductive so that these would act as high impedance to the high frequency. the detector cannot be used close to equipment. discharges occurring in the transformer and the choke are short circuited through the coupling capacitor and do not affect the measurement.

Measurement of High Voltages J. It is possible to distinguish between several types of discharges from the nature of the output displayed on the oscilloscope. The time base is produced from a phase shifting R-C network. .R. Lucas 294 Calibration is done by injection of a known step voltage into the system. The output of the amplifier is displayed on a oscilloscope having an elliptical time base. The discharge detector input circuit is shown in figure 4.43. This gives direct calibration of discharge amplitude and takes into account the response of the amplifier.

Lucas co-axial matching4.Discharge Rm Cx Test Figure b Amplif cable unit volta Displays on the oscilloscope for some typical 295 ier ge discharges are shown in figure 4.R. .Measurement of High Voltages J.44 together with corresponding waveforms arising out of external discharges as well as from contact noise. .

Lucas 296 p C p marker negative peak n m P ( Figure 4.44 Displays for typical dicharges u u e a o d c a b l l g r s ) s s a k i e b e t r i a v t e i o p n e a k .R.Measurement of High Voltages J.

but have opposite polarity. (c) External discharges: Corona produces a very symmetrical display about the negative voltage peak and as the voltage increases the discharges spread over a larger part of the ellipse but remain symmetrical. Lucas (a) For a typical oil-impregnated paper capacitor: The discharges are approximately equal in magnitude and number in the two half cycles.Measurement of High Voltages J. (d) Contact noise: Bad contacts in the system produce many small discharges at the current peaks. 297 .R. (b) For a polythene insulated cable: The discharges show the asymmetry typical of discharges between a conductor and the solid insulation for a polythene insulated cable.

Knowing the characteristics of the amplifier we can calculate the output from the circuit. The height of the pulse can be measured. Knowing the voltage sensitivity of the scope.Generation of Cfrom Figure pulses plat yfrom time base Rfor elliptical h. RX CY suppl 4. we can obtain the high frequency pulse superposed on the ellipse. we can find the magnitude. .f. By applying the output from the high frequency amplifier to the Y-plates. Lucas Oscilloscope connections for elliptical time base The oscilloscope X and Y plates are supplied from a separate source so as to form an ellipse on the screen.R.am ellipt O es discha plifier ical rge time base 298 . Then deriving a relation between the discharge from the sample and the output across the impedance we can know the discharge from the sample.Measurement of High Voltages J.

Measurement of High Voltages J.R. Lucas Calculation of internal discharge from measurements 299 .

Measurement of High Voltages J. Lucas C V vs e≡dielectric sample with void F oid qv ip u g i v u al r e e n t 4 ci . r c u E it q 300 u i v a l e .R.

c u Dielectric can be considered as being composed of a number of capacitances. Figure 4.R.t c iThe internal discharges can be analysed by considering a single flaw in rthe dielectric as shown in the figure 4.46. it Between the two electrodes (other than in the strip containing the flaw). the material is homogeneous and can be represented by a single capacitance o fbetween the electrodes.47 V c b a d a c iEquivalent circuit e l e c t r i c Measurement of High Voltages J. Lucas 301 .

d Measurement of High Voltages J.it h v The strip containing the flaw can also be considered as made up of three capacitances in series.capacity of rest of dielectric b .capacity of cavity . a .capacity of section of dielectric in series with cavity c . one representing the capacitance of the flaw and the other o itwo representing the capacitance on either side of the flaw.47.R. Lucas 302 The series capacitance on either side can be combined together to form a single capacitance as shown in figure 4.

Discharge V ∆ t v Figure V voltage V waveforms across void c Vc ac ′ If the voltage across the cavity is greater than a certain critical value. .48 . Lucas 303 Critical 4.R. and the voltage across the cavity would fall to zero. then the cavity would breakdown. the cavity capacitor discharges instantly.Measurement of High Voltages J.

R. Lucas 304 .Measurement of High Voltages J.

Lucas V Fi∆(a) 1 wh ere Δu me asu red vol tag e acr oss im pe da nce Z ΔV crit ical vol tag e Z k ∆ Cu a V V b ∆ 305 (b) .R.Measurement of High Voltages J.

R. This gives rise to a series of high frequency pulses (each of duration of the order of 100 ns).oss cav ity Measurement of High Voltages J. Thus . Consider the case of the impedance Z being a capacitor C. Lucas ΔV 1 The cavity capacity again charges up within a very short period. and again collapses discharging the charge. 306 Vo lta ge acr oss sa mp le This process repeats itself until the voltage across the cavity falls below the critical value. Figure 4.49 shows the actual circuit with the sample replaced by its equivalent circuit. The voltage ΔV1 would also be the voltage across the series combination of C and k.

∆ V = (b + a) (1 + C/k) + C In this expression b. q ∆u = (apparatuscapacitanc (1 + C / k) + C e) If the impedance across which the voltage is measured is a parallel combination of a capacitance C and a resistance R.b k ∴ ∆u = . Ck k+C b+a+ k+C b.Measurement of High Voltages J. Also. Lucas 307 ∆V1 b ∆u k = . so that = Ck ∆V ∆ V1 k + C b+a+ C+k ∆ V.ΔV is the charge dissipated in the discharge. since the cavity is small. its capacity has negligible effect on the total capacitance.R. then the above .

R. Lucas calculated value of voltage would correspond to the value before the capacitor C discharges through the resistance R exponentially. and the actual expression would be ∆u = q .Measurement of High Voltages J.e (apparatuscapacitanc (1 + C / k) + C e) t CR 308 .

If there are external discharges other than from the sample. the value of Δu would be due to the total discharge and the calculations would be in error.50 u with C & parallel R & L with C & parallel R All the apparatus other than the sample should be as discharge free as possible.Output with C 4. Lucas 309 ΔFigure only . .Measurement of High Voltages J.R.

R. Lucas A method of avoiding external discharges is by having a bridge type of circuit as shown in figure 4. from amplificatio sample n and detection 310 .51.51 . k high sup Dischar sam pass external 4.Circuit to avoid Figure ply ges ple discharges filteration.Measurement of High Voltages J.

Measurement of High Voltages J. it would measure only the discharges due to the internal flaws in the sample. so that if the detection is done across the two resistances. Lucas In this circuit.R. external discharges would affect both the resistances equally. 311 .

Test cover Cell electr s m test gap ode cell 312 .R. Lucas 4.5.52 perspex bras brass 3m .4 Measurement of dielectric constant and dissipation factor of a liquid dielectric at high frequencies using a resonance method Figure 4.Measurement of High Voltages J.

In the circuit. Since electrodes are near each other. stray capacitances have to be considered as well. 313 . Lucas Test cell used in the measurement consists of a brass cell inside which is suspended a brass electrode from a perspex cover. The test cell is connected in parallel with a variable capacitor and made part of a resonant circuit as shown in figure 4.Measurement of High Voltages J.R.53. The outer cell is the earthed electrode. R is a high series resistance used to keep the total current in the circuit very nearly constant. and there is a gap of 3 mm all round between this and the inner brass electrode. The stray capacitance C0 of the test cell can be obtained by removing the inner electrode of the test cell and with the empty cell resonance obtained.

Measurement of High Voltages J.R. Lucas 314 .

53 .Equivalent L Cv circuit for case (i) r o o . L Cv source cell r Figure 4.Test Circuit coil V a.54 .Measurement of High Voltages J.R.c. Lucas 315 Figure testR 4.

at the angular frequency ω. Otherwise the stray capacitance can be eliminated using the following procedure at the selected frequency (say 1 MHz). then ω 2 L ( C v + C0 )= 1 316 The above calculation is required only if the stray capacitance value is actually required.Measurement of High Voltages J. Lucas If Cv is the value of the variable capacitor at resonance. .R.

the variable capacitor Cv0 is varied until resonance is obtained. The test cell permits replacement of the 3 mm air gap by the insulating liquid permitting comparison of capacitances to determine the dielectric constant through the measurement of capacitance change at resonance. is Cv0 + C0. and the total capacitance will be at resonance with the coil inductance L. only the stray capacitance C0 is present. in this case. The effective capacitance. Lucas (i) With the outer cell and with only the brass screw and the perspex cover of the inner cell in position.R.Measurement of High Voltages J. The Q-factor of the circuit will be dependant on the resistance r of the coil. Under this condition. The circuit is then de-tuned to the half-power points V-mC C / C+ m √2 317 .

and C ∆ C + = ∆ C .∆ C . Usually Q is high.R.Measurement of High Voltages J.) C+ + C Q= = ∆ C+ + ∆ C C+ .C - 318 where C+ and ∆ C– are the variations at the half power points. The variable capacitance is varied in either direction from resonance until the half-power points (voltage corresponding to 1/√2) are reached. Lucas If C+ and C– are the values at the half power points (voltage corresponding to 1/√2). . then it can be shown that the Q factor at resonance can be obtained from 2 C + ( ∆ C+ .= ∆ C . so that Q = ∆ Cv The Q-factor can be determined from the half-power points.

R. Figure 4. and the circuit is again adjusted for resonance at the same frequency.Measurement of High Voltages J.55 .Equivalent R C for case (ii) C L circuit r a1 v a o 319 . Lucas (ii) The inner electrode is now screwed in.

due to the additional parallel resistance. then the total value of the capacitance required must remain the same. and Ra is the equivalent shunt resistance of the circuit with air as dielectric.Measurement of High Voltages J. This is true for all cases. then it is seen that the overall Q factor Qa is given as the parallel equivalent of the Q-factors of the coil resistance and the resistance Ra. Thus we have C v 0 + C 0 = C v1 + C 0 + C a ∴ C a = C v 0 . If the parallel equivalent resistance of the inductor is considered. so that 320 . Lucas If Ca is the capacitance of the active portion of the test cell with air as dielectric.C v1 The Q-factor of the circuit however will be different from the earlier value.R. The Q-factor corresponding to the resistance Ra is ωCRa.

.R.Measurement of High Voltages J. Lucas 321 1 1 1 = + ω C Ra Qa QL (iii) The liquid is now introduced into the test cell.

. and k is the relative permittivity of the liquid dielectric. so that the surface condition of the perspex is not changed. then the capacitance of the active portion of the test cell with the liquid would be kCa. Lucas 322 Figure R C 4.R.Measurement of High Voltages J.] If Rk is the equivalent shunt resistance of the liquid.Equivalent C L circuit for case (iii) a k k v 2 o r [The liquid level should be slightly below the perspex cover.56 .

R.Cv 2 Cv 0 .Cv 1 323 Also we have the equivalent Q factor Qk equivalent to the parallel equivalent.Measurement of High Voltages J. then C v0 + C0 = C v2 + C0 + k C a giving k Ca = Cv 0 . Lucas If Cv2 is the value of the variable capacitor at resonance. Thus 1 1 1 1 = + + ω C Ra ω C Rk Qk QL Thus the inverse of ωCRk can be determined from .Cv 2 ∴ k = Cv 0 .

= C C Qk Qa Qk Qa loss factor = 1 ω Ck R k = 1 C .Measurement of High Voltages J. Lucas 324 1 1 1 = ω C Rk Qk Qa 1 1 1 ( ∆ C ) k 1 ( ∆ C )a . [ ∆ C k .R.∆ Ca ] = . = .∆ Ca ] k Ca C The loss factor of the dielectric is given by . can be calculatedusing = . k Ca  Q k Q a  k Ca C 1 = . ω C R k Ck 1 1 C [ ∆ C k . .

.Cv 2 Note: in making connections it is essential that care is taken to minimise stray capacitances by using short leads. and the components should not be disturbed during the experiment.∆ Ca i . Lucas 325 ∆ C k . e .R.Measurement of High Voltages J. loss factor = Cv 0 .

R.57 .5.5 Ionic Wind Voltmeter safety hot h.Measurement of High Voltages J. eart G gap wire 326 line hFigure 4.v. Lucas 4.Ionic Wind .

of platinum-gold alloy. This is referred to as electric wind and is brought about by repulsion of ions from surface of the point by the intense electro-static field. A similar wind is observed also at the earth electrode. Before the high voltage is applied. a hot wire. 327 . Lucas When a highly charged point is situated in air or other gas. the electric wind cools the hot wire and hence reduces the resistance. In the ionic wind voltmeter.Measurement of High Voltages J. These ions colliding with uncharged molecules of air carry them with it setting up the electric wind. the bridge is balanced.R. a movement of the air around the point is observed. included in one arm of a Wheatstone bridge network is used as the earthed electrode of high-tension gap. When the voltage of the gap exceeds the "threshold voltage" (voltage required before the potential gradient is sufficient for ionization to commence).

value of alternating voltages and direct voltages. Lucas This reduction causes an appreciable out-of-balance voltage in the bridge.v.6 Dumb-bell Voltmeter shiel ellips F d oid 328 . The voltmeter can be used to determine either the peak value or the r. and the instrument is calibrated for a sine wave.Measurement of High Voltages J.s. and the robust construction and freedom from disturbances by temperature and weather conditions which make it suitable for outdoor use.m. may be measured by an observer at some distance from the charged conductors. The principle advantages are that the h.5. 4.R. The voltage waveform influences the instrument reading.

Bruce in which the period of oscillation of a conducting spheroid in an electro-static field was determined.M.R.Measurement of High Voltages J.s.05%. voltage was developed by F. with an accuracy of 0. Lucas A rather specialized way of measuring r. The instrument is shown on figure 4. This enabled the voltage to be determined in terms of length and time.58.m. 329 .

5 High Voltage Generators for Testing .

Thus special methods may be used which are not applicable when generating high voltage in high power applications. and secondly for testing of equipment used in power transmission. Therefore high voltage testing does not usually require high power. high voltages are generated for testing of insulation. the primary source of power is at low voltage (400 V three phase or 230 V single phase.0Generation of High Voltages Generation of High Voltages for Testing Power systems engineers are interested in high voltages primarily for power transmission. Generation has to be carried out in the testing laboratory. In this chapter. In many testing laboratories. at 50 Hz). impedances involved are extremely high (~ MΩ) and the currents small (<1 A).331 J R Lucas 5. . Since insulation is usually being tested.

Generation of High Voltages for Testing . a cascade arrangement of several transformers is used. for high voltages to reduce the cost (insulation cost increases rapidly with voltage) and make transportation easier.1 Cascade arrangement of transformers Figure 5.1 shows a typical cascade arrangement of transformers used to obtain up to 300 kV from three units each rated at 100 kV insulation.1. The low voltage winding is connected to the primary of the first transformer. 5.1 Generation of High Alternating Voltages Single transformer test units are made for high alternating voltages up to about 200 kV.332 J R Lucas 5. One end of the high voltage winding is also earthed through the tank. and this is connected to the transformer tank which is earthed. However.

333 J R Lucas The high voltage end and a tapping near this end is taken out at the top of the transformer through a bushing. Generation of High Voltages for Testing . and forms the primary of the second transformer.

334 J R Lucas Generation of High Voltages for Testing 9 E9 bush 300 200 199 100 99 hv 1Figure Insulati ing kVng output Insulati Pedesta ng l Pedesta l 5.Cascade arrangement of transformers . .

a n The actual arrangement could be different for practical reasons. 335 J R Lucas Generation of High Voltages for Testing . a r What is shown in the cascade transformer arrangement is the basic r principle involved.I n t h One end of this winding is connected to the tank of the second e transformer to maintain the tank at high voltage. g e m e n t s h o w n . c a This cascade arrangement can be continued further if a still higher d e voltage is required. c a The secondary of this transformer too has one end connected to the tank s and at the other end the next cascaded transformer is fed.

The current i would be given by . - Resonance circuit 5.1. Let R represent the equivalent parallel resistance across the coil and the device under test.336 J R Lucas Generation of High Voltages for Testing V E L C R Figure 5.2 Resonant Transformers The resonance principle of a series tuned L-C circuit can be made use of to obtain a higher voltage with a given transformer.

337 J R Lucas Generation of High Voltages for Testing 1 jω L R + jω C R+ jω L jω L R so that v = i .ω L C R jω L i= E Since R is usually very large. v = = at resonance 2 R + jω L . the Q factor of the circuit (Q = R/Lω) would be very large.Q Lω It can thus be seen that a much larger value that the input can be obtained across the device under test in the resonant principle.e. and the output voltage would be given by R |v| = E . . E E. = E .R i . R + jω L -ω2 L C R .

338 J R Lucas Generation of High Voltages for Testing 50 Hz L Test Cair-cored Figure air-cored transformer supply R coil device 5.resonant transformer . .

This would result in a smaller transformer having fewer turns.339 J R Lucas Generation of High Voltages for Testing at resonance ω = 2π f = 1 LC Figure 5. it is an advantage to select a frequency higher than power frequency (50 Hz). and the following is sometimes used to obtain a supply at three times power frequency. particularly when the final requirement is a direct voltage.3 shows the application of the resonance principle at power frequency. For certain applications. . High voltage high frequency voltages are not readily available. and also simplifies the smoothing after rectification.

Figure 5.4 shows the circuit arrangement. no power frequency voltage would remain and only the third harmonic component would be present.340 J R Lucas It makes use of the fact that the magnetising current of a transformer has a high third harmonic component. Thus if an open delta secondary is used. Generation of High Voltages for Testing .

Resonant transformer 3 Figure 5.341 J R Lucas Generation of High Voltages for Testing L rd C 3-High . voltage harmon phase rd ic 3 50 compo harmonic nent componen Hz t output suppl y Air-cored coils are used to simplify the construction and the insulation. .

1. The damped oscillations are obtained by the use of a Tesla coil. The high voltage secondary coil has a large number of turns. or in some cases. The importance of testing with high frequency is that high frequency oscillations cause failure of insulator at comparatively low voltage due to high dielectric los and consequent heating. The tesla coil constitutes the high voltage transformer.3 High frequency high voltages High frequency (few kHz to Mhz) high voltages are required in testing apparatus for behaviour with switching surges. Thus it is necessary to produce damped high frequency voltages. Generation of High Voltages for Testing . oil. together with a circuit containing a quenched spark gap. It consists of two air-cored coils which are placed concentrically. insulation flashover etc. and is wound on a frame of insulating material. the insulation between turns being air.342 J R Lucas 5.

Tesla coil circuit for trig equip Clow L ironTesla Figure ring ger ment voltag 1 core 2coil sphere under gap e transfo gap Hz test 50 rmer Generation of High Voltages for Testing The supply is usually 50 Hz to the primary of the high voltage testing transformer.343 J R Lucas The primary winding has only a few turns wound on an insulating frame. . [In the circuit shown. Measu 5. .] The primary circuit of the tesla transformer also contains a trigger spark gap. C2 includes the capacitance of the sphere gap used for measurement.

Thus there will be a hundred of these trains of damped oscillations per second. This charge and discharge of capacitor C1 takes place twice in one voltage cycle. During the time taken for this train of oscillations to die away. At this voltage. the capacitor C1 is charged up to some maximum voltage. and a train of damped oscillations of high frequency is produced in the circuit containing C1.344 J R Lucas Since the supply to the primary of the tesla transformer is alternating. the spark gap and the primary winding of the tesla transformer. the spark gap is conducting. Generation of High Voltages for Testing . the capacitor C1 discharges. the trigger gap breaks down. which depends upon the secondary side of the supply transformer. due to the formation of an arc across it. and upon the setting of the trigger gap.

The circuit parameters are generally such that the resonant frequencies of the two sides are the same. its actual value depending upon the inductance and capacitance of the oscillatory circuit.α 2 t + C e.α 3 t + D e.345 J R Lucas The frequency of oscillations themselves is very high (about 100 kHz usually).α 1 t + B e. C 1 L1 = C 2 L 2 ≈ 1 2 Generation of High Voltages for Testing ω The expression for the voltage variation being obtained as the solution to a fourth order differential equation. v = A e.α 4 t . whereα is complex The solution to the differential equation will generally be in conjugate pairs. .

. φ 1 . ω1 . then the damped resonant frequencies are nearly equal (ω1 ≈ ω2). etc Thus the solution can be written in the form v = A1 e..j ω .a2 t sin ( ω 2 t + φ 2 ) where a1 .346 J R Lucas Generation of High Voltages for Testing α 1 = a + j ω . a 2 . . A1 ... A 2 . α 2 = a .a1 t sin ( ω 1 t + φ 1 ) + A2 e. φ 2 are constants If the two undamped frequencies are equal (corresponding to L1C1 = L1C1). ω 2 . The exponential decays of the components of the voltage depends on the resistance values.

and the other with a frequency equal to the difference frequency between the primary and the secondary resonance frequencies. so that the voltage of primary is minimum when the secondary voltage is maximum and vice versa.φ 2 ) sin ( ω 1 t + φ 1 ) + sin ( ω 2 t + φ 2 ) = 2 sin . the above result will be modified by the constants A1 and A2.ω 2 .t +φ1 .t +φ1+φ 2 ) ( ω1 . If the magnitudes and decays were not considered equal. and the decays also equal. . then (ω1 + ω2)/2 ≈ ω. cos 2 2 Generation of High Voltages for Testing If ω1  ω2.347 J R Lucas If amplitudes A1 and A2 are equal. one of which is of very nearly the resonant frequency. then the summation in v would have the form ( ω1 +ω 2 . so that the sum of the two sine terms represents a product of terms. -a1 t -a2 t and the exponential decays e and e . The energy tends to get transferred from primary to the secondary and vice versa.

V o l .348 J R Lucas Generation of High Voltages for Testing (b) (a) Secondary voltage Primary voltage F i g u r e 5 .

6 (b) shows the corresponding waveforms s with quenching. r Figure 5. a c r o s s 349 J R Lucas Generation of High Voltages for Testing . a This can be done by preventing the energy from travelling backwards and v forwards in the tesla transformer by quenching the trigger gap by air blast e cooling. f When the primary voltage is zero. w What we require is a single series of short duration pulses.6 (a) shows the primary and secondary voltage waveforms m without quenching and figure 5.a g e Oscillation would occur which would be damped out due to the resistance in the circuit. the blast of air removes the spark in the o primary gap so that the energy is confined to the secondary.

t e 5. t 5. a low pass filter may be used to smooth the output. o In r theory.2. r .2Generation of High Direct Voltages s Generation of high direct voltages are required in the testing of high l voltage direct current apparatus as well as in testing the insulation of a cables and capacitors where the use of alternating voltage test sets become 350 J R Lucas Generation of High Voltages for Testing impractical due to the steady high charging currents. s The rectifiers used must be high voltage rectifiers with a peak inverse f voltage of at least twice the peak value of the alternating voltage supply.1 Rectifier circuits r One of the simplest methods of producing high direct voltages for testing a is n to use either a half-wave for full-wave rectifier circuit with a high alternating voltage source. however when the test device is highly capacitive no smoothing is required. m Even otherwise only a capacitance may be used across the test device for e smoothing.

automatic discharging is provided which discharges the capacitors to earth. .2 Voltage Multiplier Circuits Both full-wave as well as half-wave circuits can produce a maximum direct voltage corresponding to the peak value of the alternating voltage. .2. When higher voltages are required voltage multiplier circuits are used. In certain test sets.Half-wave and fullR Rectif voltag full-wave rectifier ier (b) L e 50 Hz Generation of High Voltages for Testing In testing with high voltage direct current care must be taken to discharge any capacitors that may be present before changing connections.351 J R Lucas Figure 5.7 shows the half-wave and the full wave arrangements. 5. Figure B A – +(a) half-wave rectifier Clow 5.

352 J R Lucas Generation of High Voltages for Testing l. . the transformer will be of small rating that for the same direct voltage rating with only simple rectification. Voltage Doubler Circuit The voltage doubler circuit makes use of the positive and the negative half cycles to charge two different capacitors. Figure 5.v.8 shows a voltage doubler circuit. In this case. These are then connected in series aiding to obtain double the direct voltage output. – 2V Vmx + Figure 5.0 Voltage supa sin+ – mx ωt a ply The common circuits are the voltage double circuit and the CockroftWalton Circuit. .

The circuit is shown in figure 5. Cockroft-Walton Circuit When more than doubling of the voltage is required.9. Generation of High Voltages for Testing .353 J R Lucas Further for the same direct voltage output the peak inverse voltage of the diodes will be halved. the Cockroft-Walton voltage multiplier circuit is commonly used.

Cockroft-Walton Circuit supply Let Vmax be peak value of secondary voltage of high voltage transformer.354 J R Lucas Generation of High Voltages for Testing Figure V + d c bm a3 01 D4 a. consider that charging of capacitors actually takes place stage by stage rather than somewhat simultaneously. .c. . C2ax EHT 5. To analyze the behaviour.

Generation of High Voltages for Testing . and the voltage at b would be zero. the point a will be at 2 Vmax. the capacitor C1. and the secondary winding. diode D1 is reverse biassed. During the following cycles.355 J R Lucas This assumption will not invalidate the result but will make analysis easier to follow. the capacitor C1 charge up to voltage Vmax. capacitor C2 would be uncharged. Consider the first part of the circuit containing the diode D1. Initially. capacitor C1 will not discharge (or will not charge up in the other direction) and the peak of this half cycle. depending on whether the secondary voltage and the capacitor voltage are opposing or assisting. Since during positive half cycle which follows. the potential at a will vary between 0 and 2 Vmax. During the first negative half cycle of the applied voltage.

and the capacitor C2 would charge to 2 Vmax. Thus once C2 has charged up. As the voltage at a again increases to 2 Vmax. as C3 has not discharged. Once the voltage at b has reached 2 Vmax. the diode D2 is forward biassed. the voltage at c increases. this diode too would be reverse biassed and the capacitor C2 would not discharge. the voltage at a would be less than or equal to the voltage at b. the diode D3 would initially be forward biassed for almost the whole cycle. C3 is also initially assumed uncharged. The voltage at b would now remain constant at 2 Vmax. Generation of High Voltages for Testing . and thus the diode D3 is reverse biassed and C3 would not discharge. Since the voltage at a varies between 0 and 2 Vmax. As a reaches 2Vmax voltage at c rises to 4Vmax. Capacitor C3 charges until it reaches 2 Vmax when b is 2 Vmax and a is 0.356 J R Lucas Thus as the voltage at a varies between 0 and 2 Vmax.

1. the voltage at c varies between 2 Vmax and 4 Vmax. to 2 Vmax).357 J R Lucas Thus after charging up has taken place. Generation of High Voltages for Testing . since the voltage at b is a constant at 2 Vmax and the voltage at c varies between 2 Vmax and 4 Vmax initially. the diode D4 is forward biassed and C4 charges up to the maximum difference between d and b (i. during most of the cycle.e. This sequence of voltages gained is shown in Table 5. Thus once the capacitors are charged up the voltage at d remains constant at 4 Vmax. This occurs when the voltage at c is 4 Vmax and the voltage at d would now be 4 Vmax. since there is no discharge path. Assuming C4 also to be initially uncharged. since the capacitor C4 has charged up it would not discharge. As the voltage at c falls from 4 Vmax to 2 Vmax.

. or when it is loaded. a current is drawn from the generator.358 J R Lucas Generation of High Voltages for Testing Cycle Location a b c d 0 – 0 0 0 0 T/2 + 2Vm T – 0 3T/2 + 2Vm 2T – 0 5T/2 + 2Vm 3T – 0 7T/2 + 2Vm 4T – 0 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm 2Vm 0 0 2Vm 4Vm 2Vm 4Vm 2Vm 4Vm 2Vm 0 4Vm 4Vm 4Vm 4Vm 4Vm 4Vm Table 1 When the generator is used for a test. and the voltage falls slightly depending on the load. and the capacitors lose some of their charge to the load.

359 J R Lucas As the voltage across any of the capacitors drops. Van de Graeff generator Generation of High Voltages for Testing .3 Electrostatic generators Electrostatic generators using the principle of charge transfer can give very high direct voltages. then at some point in the applied alternating voltage cycle. the corresponding diode would become forward biassed and charging up of the capacitor would once again result. 5. there would be a small ripple in the output voltage. The basic principle involved is that the charge is placed on a carrier. and raised to the required potential by being mechanically moved through the electrostatic field.2. either insulating or an isolated conductor. Thus when a load is connected.

positivespray corona −Figure •insulating ++driver − device − belt + pulley − − + − + + − − − + − − + − − + − − + + + + + + Van de Graeff Generator .360 J R Lucas Generation of High Voltages for Testing +−moving 5.

with corona discharge taking place at the positive end of the source.361 J R Lucas The Van de Graeff generator is one of the methods used to obtain very high voltages. and their use is restricted to low current applications. Generation of High Voltages for Testing . However they cannot supply much currents and the power output is restricted to a few kilowatt. The corona formation (spray) is caused by a core like structure with sharp points (corona spray device). The Van de Graeff generator uses an insulating belt as the carrier of charge. The generator consists of a low direct voltage source.

The upper electrode at which the charge is collected has a high radius of curvature and the edges should be curved so as to have no loss. a smaller capacitance gives a larger voltage. Q V = C Generation of High Voltages for Testing .362 J R Lucas Charge is sprayed onto the belt at the bottom by corona discharges at a potential of 10 to 100 kV above earth and carried to the top of the column and deposited at a collector. The generator is usually enclosed in an earthed metallic cylindrical vessel and is operated under pressure or in vacuum. The upper electrode has a smaller capacitance to earth on account of the larger spacing involved. The higher voltage of the upper electrode arises from the fact that for the same charge.

. With a single source at the lower end.363 J R Lucas Generation of High Voltages for Testing The potential of the high voltage electrode rises at a rate of dV 1 dQ I = = dt C dt C where I is the net chargingcurrent A steady potential will be reached by the high voltage electrode when the leakage currents na the load current are equal to the charging current. the belt moves upwards with a positive charge and returns uncharged. The edges of the upper electrode are so rounded as to avoid corona and other local discharges.

Generation of High Voltages for Testing .364 J R Lucas Charging can be made more effective by having an additional charge of opposite polarity sprayed onto the belt by a self inducing arrangement (negative corona spray) using an ingenious method. This arrangement effectively doubles the charging rate.

+ + + + + .365 J R Lucas Sames Generator Rotor Exciter H. F i g u r -e5 ..T.-+ + + Generation of High Voltages for Testing + + + + + + + + + + + + .-S a -m e s G e .

For example. o 2 pole generator is shown in figure 5. 366 J R Lucas Generation of High Voltages for Testing . the rotor is a hollow cylinder made of an insulating material. a tIn this. In the Sames generator. Electric charges are deposited on the surface of the rotor which is driven by an electric motor to effect the transfer of charges in the field. The whole unit is sealed in a pressure unit and insulated ith hydrogen at a pressure of 10 to 25 atmospheres. In this the power output will depend on the size of rotor. The number of poles will determine the current and the voltage. but other number of poles are A ralso possible. the charge is carried on the surface of an insulating cylinder.11. a four pole rotor will produce twice the current at half the voltage of that of a two pole machine of the same size.e r This is a more recent form of the electrostatic generator.

367 J R Lucas Generation of High Voltages for Testing .

6 High Voltage Testing .

dielectric loss per unit volume. the power factor or the total dielectric loss.369 J R Lucas High Voltage Testing 6. . The first two can be measured using the High Voltage Schering Bridge. The tests carried out on completed equipment are the measurement of capacitance. Thus by suitable testing procedure we must ensure that this is done. The tests carried out on samples of dielectric consist generally of the measurement of permittivity. High voltage testing can be broadly classified into testing of insulating materials (samples of dielectrics) and tests on completed equipment. the ultimate breakdown voltage and the flash-over voltage. and the dielectric strength of the material.0High Voltage Testing Procedure Electrical equipment must be capable of withstanding overvoltages during operation.

.370 J R Lucas High Voltage Testing The breakdown voltage tests on completed equipment is only done on a few samples since it permanently damages and destroys the equipment from further use. However since all equipment have to stand up to a certain voltage without damage under operating conditions. all equipment are subjected to withstand tests on which the voltage applied is about twice the normal voltage. but which is less than the breakdown voltage.

and for work tests on high voltage transformers. These tests are made upon specimens of insulation materials for the determination of dielectric strength and dielectric loss. for routine testing of supply mains.1General tests carried out on High voltage equipment 6.1.371 J R Lucas High Voltage Testing 6. voltages as high as 2000 kV may be used.1 Sustained low-frequency tests Sustained low frequency tests are done at power frequency (50 Hz). For testing of porcelain insulators and in high tension cables. . Since the dielectric loss is sensitive to electric stress. and are the commonest of all tests. porcelain insulators and other apparatus. the tests are carried out at the highest ultimate stress possible.

The high voltage is applied across the device under test by means of a transformer. the size of the transformer and the insulation bushing necessary may be reduced in size.c.372 J R Lucas High Voltage Testing High voltage a. . A 230 V piece of equipment may thus be subjected to about 1. Each one of these devices are subjected to a high voltage of 1 kV + 2 × (working voltage).5 to 2 kV. The transformer need not have a high power rating. By means of cascading. If a very high voltage is required. the transformer is usually build up in stages by cascading. tests at 50 Hz are carried out as Routine tests on low voltage (230 or 400 V) equipment. These tests are generally carried out after manufacture before installation.

the terminal voltage would drop due to the high current caused. A resistance of about 1 ohm/volt is used in series with the transformer so as to limit the current in the event of a breakdown to about 1 A. .c. . but would be a simple device) such as a tube filled with water. ice/ kV kΩ Ω generation test circuit und V er test The resistance used could be of electrolyte type (which would be far from constant. Dev 200 1 Figure 6.a.373 J R Lucas High Voltage Testing The transformers are usually designed to have poor regulation so that if the device under test is faulty and breakdown occurs.

There would be switches that would automatically be operated when the door to the area is opened etc. .374 J R Lucas High Voltage Testing In all high voltage tests. safety precautions are taken so as to ensure that there is no access to the testing area when the high voltage is on.

a high charging current would be drawn and the transformer used would have to have a current rating.1.2 High Voltage direct current tests These tests are done on apparatus expected to operate under direct voltage conditions. conditions. . conditions.c. due to the inconvenience of the use of high capacity transformers required for extra high tension alternating voltage tests and due to transport difficulties.5 hrs. but before energising it ) to carry out a high voltage test under d.c. The test voltage would be about 2 (working voltage ) and the voltage is maintained from 15 min to 1. and also where. conditions. If the tests are done under a. It is thus normal to subject the cable (soon after laying it. test is the testing of cables which are expected to operate under a.c.375 J R Lucas High Voltage Testing 6. A special feature of importance of the d. alternating voltage tests cannot be performed after installation.c.

c test is not complete equivalent to the corresponding a. it is the layers of different dielectrics that determine the voltage distribution in the cable. conditions . . it is likely that the cable will stand up to the required a. voltages have already been described.c. while in the a. it is the leakage resistance which would determine the voltage distribution.c. conditions.c. Although the electric field differs in the 2 cases. voltage. The methods used to generate these high d.376 J R Lucas High Voltage Testing This d.c.

1. breakdown or flashover occurs in most cases as a result of high frequency disturbances in the line.3 High-frequency tests High frequency tests at frequencies varying from several kHz are important where there is a possibility of high voltage in the lines etc. High voltage tests at high frequency are made at the manufacturing works so as to obtain a design of insulator which will satisfactorily withstand all conditions of service. and in insulators which are expected to carry high frequency such as radio transmitting stations. .. Also in the case of porcelain insulators.377 J R Lucas High Voltage Testing 6. It is also found that high frequency oscillations cause failure of insulation at a comparatively low voltage due to high dielectric loss and heating. these being due to either switching operations or external causes.

378 J R Lucas High Voltage Testing F .

Sudden interruptions in the line would give rise to resonant effects in the line which would give rise to voltage waves in the line of high frequency. it is possible that breakdown or flash over would occur due to high frequency over voltages produced by faults or switching operations in the line.379 J R Lucas High Voltage Testing In the case of power line suspension insulators. The behaviour of insulating materials at high frequencies are quite different to that at ordinary power frequency. The movements of charge carriers would be different. These might cause flashover of the insulators. The dielectric loss per cycle is very nearly constant so that at high frequencies the dielectric loss is much higher and the higher loss causes heating effects.2). . so that they may go about half–way and turn back (figure 6. At high frequency the polarity of electrodes might have changed before the charge carriers have travelled from one electrode to the other.

6. (a) Tests with apparatus which produces undamped high–frequency oscillations. high frequency transients occur whose frequency depends on the capacitance and inductance of the line and will be about 50 kHZ to about 200 kHZ. (b) Tests with apparatus producing damped high–frequency oscillations. These are damped out with time.1. Undamped oscillations do not occur in power systems.380 J R Lucas High Voltage Testing There are two kind of high frequency tests carried out. breakdown of insulators and of the end turns of transformer connections to line. but are useful for insulation testing purposes especially for insulation to be in radio work. . When faults to earth or sudden switching of transmission lines occur.4 Surge or impulse tests These tests are carried out in order to investigate the influence of surges in transmission lines.

conditions occurring on the flash over to line are simulated. Overvoltages of much higher duration also arise due to line faults.381 J R Lucas High Voltage Testing In impulse testing. By the use of spark gaps. to represent surges generated due to lightning. for which impulse waves such as 100/5000 μs duration may be used. In surge tests it is required to apply to the circuit or apparatus under test.2/50 μs wave is generally used. switching operations etc. the IEC Standard impulse wave of 1. although the total duration of the lightning stroke may be a few seconds. The total duration of a single lightning strike os about 100 μs. a high direct voltage whose value rises from zero to maximum in a very short time and dies away again comparatively slowly. Methods of generating such voltages have already been discussed earlier. .

in order to ensure that their finished products will give satisfactory performance in service . the most general tests upon insulating materials are carried out at power frequencies . .382 J R Lucas High Voltage Testing While impulse and high frequency tests are carried out by manufacturers .

The flash–over is due to a breakdown of air at the insulator surface. As the flash–over under wet conditions and dry conditions differ .383 J R Lucas High Voltage Testing Flash–over Tests Porcelain insulators are designed so that spark over occurs at a lower voltage than puncture. thus safeguarding the insulator. using an impulse generator delivering a positive 1/50 μs impulse wave. tests such as the one minute dry flash–over test and the one minute wet flash–over test are performance. The voltage shall be increased to the 50 percent impulse flash–over voltage (the voltage at which approximately half of the impulses applied cause flash– over of the insulator) . (i) 50 percent dry impulse flash–over test. in service against destruction in the case of line disturbances. and is independent of the material of the insulator. Flash–over tests are very importance in this case .

but rather the voltage for puncture of a given thickness ( ie.000 and 11. dielectric strength ).) is applied . it is not the voltage which produces spark–over breakdown which is important . The voltage shall then be increased gradually until flash– over occurs . The voltage is raised to this value in approximately 10 seconds and shall be maintained for one minute. (iii) Wet flash–over and one minute rain test In this the insulator is sprayed throughout the test with artificial rain drawn from source of supply at a temperature within 10 degrees of centigrade of the ambient temperature in the neighborhood of the insulator.000 ohm cm. In the case of the testing of insulating materials .384 J R Lucas High Voltage Testing (ii) Dry flash–over and dry one–minute test In this the test voltage (given in the B. The resistivity of the water is to be between 9. .S.S.

those of dielectric strength. (h) temperature and humidity when the test is carried out (i) moisture content of the sample. (a) thickness of the sample tested (b) shape of the sample (c) previous electrical and thermal treatment of the sample (d) shape . The dielectric strength of a given material depends. . upon many factors including. the latter been intimately connected with the dielectric strength of the material. dielectric loss and power factor. material and arrangement of the electrodes (e) nature of the contact which the electrodes make with the sample (f) waveform and frequency of the applied voltage (if alternating ) (g) rate of application of the testing voltage and the time during which it is maintained at a constant value .385 J R Lucas High Voltage Testing The measurements made on insulating materials are usually. size . apart from chemical and physical properties of the material itself.

386 J R Lucas High Voltage Testing 6. If the rate of increase of heat dissipated. when an electrostatic field is applied to them. a stable condition (thermal balance) will be reached. These losses result in the formation of heat within the material. Most insulating materials are bad thermal conductors. the conductivity of such materials increases considerably with increase of temperature. is greater than the rate of increase of dielectric loss with temperature rise. and the dielectric losses.2Testing of solid dielectric materials 6. . the temperature thus building up from the small initial temperature rise. therefore.1 Nature of dielectric breakdown Dielectric losses occur in insulating materials. it is not rapidly carried away by the material. rise and produce more heat. with rise of temperature. Now.2. so that even though heat so produced is small.

The measurement of dielectric loss in insulating materials are very important. dielectric loss measurements are now generally recognized as the most reliable guide to the quality and condition of the cable. the insulation will breakdown owing to the excessive heat production which burns material. . Hence the decrease in breakdown voltage with increasing time of application.387 J R Lucas High Voltage Testing If the latter rate of increase is greater than the former. as they give a fair indication as to comparative dielectric strengths of such materials. In the case of cable. The dielectric losses per cubic centimetre in a given material and at a given temperature. increasing temperature and also the dependence voltage upon shape. size and material of electrodes and upon the form the electric field. are directly proportional to the frequency of the electric field and to the square of the field strength.

388 J R Lucas High Voltage Testing d spherical F ieelectrode le ct ri c m at e ri al .

5 cm in diameter) with a wall or partition of the material between them 0. . The electrical stress is applied to the specimen by means of the two spheres fitting into the recesses without leaving any clearance.2.5mm thick. is taken and recessed on both sides so as to accommodate the spherical electrodes (2.389 J R Lucas High Voltage Testing 6. especially at the centre. Sometimes insulators after manufacture are found to contain flaws in the form of voids or air spots.2 Determination of dielectric strength of solid dielectrics A sheet or disc of the material of not less than 10cm in diameter. The applied voltage is of approximately sine waveform at 50Hz. This voltage is commenced at about 1/3 the full value and increased rapidly to the full testing voltage.

causing physical decomposition. the life of the component is longer. polymerisations. If the applied voltage is small..which would gradually diffuse into the material the by-products. In plastic type of materials. causing chemical destruction. . and if present would gradually deteriorate and cause ultimate breakdown after a number of years.there might be carbonizations. From design considerations the voltage to be applied if a particular life span is required can be calculated. The useful life of a component using such material will depend on the weak spots and the applied voltage. High degree ionisations caused in these spots would give rise to high energy electrons which would bombard the rest of the material. chemical decomposition etc.390 J R Lucas High Voltage Testing These spots (due to non-homogeneity) have a lower breakdown strength than the material itself.

391 J R Lucas High Voltage Testing l m a F o a p s x p s i li f m e a u d c m v t o o s lt r a f g e t w a o n r k δ i n g .

since beyond this. the specimen would keep on deteriorating. As the voltage is increased at a certain value it starts increasing at a faster rate. the applied voltage can be higher than this safe value. Thus small flaws if present would not cause much of a variation in the overall p. . If the apparatus need be used only for a short period.o lt a gThe schering bridge type of measurement gives an average type of emeasurement.f. and the power loss indicates the value over 392 J R Lucas High Voltage Testing the whole of the length.where the p. In the schering bridge type of measurement such flaws would not be brought out.f. This is the long time safe working voltage. The loss factor of a material does not vary much for low voltages.

Discharges caused are made to give pulses to a high frequency amplifier. such small individual spots cannot be detected. . The method is to apply suitable high voltage to sample. make and break cycles). In high voltage transformers also there might be such small discharges occurring which would not be measured by the schering bridge. since it measures the overall loss. It is necessary that such spots are detected as these increase with time and finally cause its breakdown. Figure . and subject it to a number of duty cycles ( heat cycles.393 J R Lucas High Voltage Testing weak spots conductor dielectric 6.sample of cable showing weak spots In a long length of cable. the greater part of the cable would be in good conditions but with a few weak spots here and there. In a Schering bridge type of measurement.

6.3Impulse Testing . The methods of discussion have been discussed in an earlier chapter.394 J R Lucas High Voltage Testing The discharges caused are observed before and after such duty cycles to see whether there is any appreciable increase in the pulse intensity after the cycle of operation.

395 J R Lucas High Voltage Testing (Figure 6. .Observed impulse waveforms t ii vi i) ) .

so we have it on a high frequency ( single sweep and high speed ) oscilloscope. To test the ultimate impulse strength. The impulse test level is determined by the operating level (4 to 5 times the normal operating value ) Apply on to the sample a certain number (say 10) positive impulse and 10 negative impulses of this particular value. . breakdown of the insulator due to the application of the impulse voltage will be indicated as in (i). The damage may not be immediately visible. They should withstand this voltage without any destruction. apply increasing amounts of impulse voltage until destruction occurs.396 J R Lucas High Voltage Testing These are done as tests on sample of apparatus. during the tests it is necessary to see whether there is any damage. In the event of complete damage.

397

J R Lucas

High Voltage Testing

If the insulator has suffered only a minor damage the wave form would show no distortion , but would show as in (ii). If there is no damage caused due to the impulse, the waveform will be complete and undistorted as in (iii). In testing high voltage insulators whose actual breakdown is in air (i.e flashover takes place before breakdown of insulator), the porcelain itself can be tested by immersing whole insulator in liquid of high permeability. Thus there would be no outside flashover, and actual breakdown of the insulator would occur.

398

J R Lucas

High Voltage Testing

exter punct liq F nal ure uid flash over

In

specifying

the

flashover characteristic in air we give the 50% flashover characteristic. This is done as flashover occur at the same voltage on each application of the impulse .

399

J R Lucas

High Voltage Testing

5 1 I P F 0 0 m r % 0 % p o %

u b b la r s b e il a t it k e y d s o t w vWe apply different values of test voltages (impulse) and the voltage at f n o fl which there is 50% probability of breakdown is taken as 50% flashover v lt a o avoltage. s lt g h a e o g v e e r

400

J R Lucas

High Voltage Testing

The impulse flashover voltage also depends on the time lag of the applied impulse before flashover time lag of the applied impulse before flashover occurs. Thus we have also got to determine the time lag characteristics for breakdown. IIf fl ti the voltage remains above a critical value long enough, flashover F aoccurs. m p s e The time lag before flashover occurs depends on the statistical time lag u h l oand on the formation time lag.
s v e Depending on the volume of space between the gap, and also depending r

on the nature of shielding, a certain time will be taken for enough free electrons to be set free. This is the statistical time lag.

401

J R Lucas

High Voltage Testing

Once the electrons appear, depending on the voltage applied, they multiply and ionise the space. once the space becomes conducting, vflashover occurs. o This is formation time lag. lt aTo determine the time lag characteristic of a device, we can use the gimpulse generator to generate impulses of gradually increasing amplitude eand determine the time of breakdown. At each value, the test must be repeated a number of times so as to obtain consistent values.

402

J R Lucas

High Voltage Testing

V t olta Figure 6. - Time lag ge

This type of characteristic is important in designing insulators.If a rod gap is to protect a transformer. Then the breakdown voltage characteristic of the rod gap must be less than that of the transformer so as to protect it. If the characteristic cross, protection will be offered only in the region where the rod gap characteristic is lower than that of the transformer.

403

J R Lucas

High Voltage Testing

System Voltage 11 kV 33 kV 66 kV 132 kV 275 kV

I.E.C. Impulse Withstand Voltage 75 kV 170 kV 325 kV 550 kV 1050 kV

In obtaining the breakdown characteristic of a transformer we do not attempt such tests that cause total destruction on transformers as they are expensive. What is done is we take a sample of the material used as insulators for the transformers and then apply these tests till puncture takes place. Thus the transformer characteristic is obtained by such tests on samples.

404

J R Lucas

High Voltage Testing

To obtain one point on the voltage vs time lag characteristic we would have to do a large number of tests and take the mean,as these values vary from sample to sample. The sample would have to be surrounded by a liquid material of high permittivity so that external flashover would not occur. The impulse test voltage recommended by I.E.C. (International Electrotechnical Commission) are given in the table 6.1. The recommendation is that device when subjected to this voltage should not suffer permanent damage or minor partial damage. The voltage is set at slightly less than the withstand voltage and gradually increase to test value. About 10 positive impulses and 10 negative impulses are applied.

6.4Voltage Distribution in a Transformer Winding
Consider the entering of an impulse voltage on the terminating transformer, as shown in figure 6.11.

405

J R Lucas

High Voltage Testing

lightni voltag surge transformer Fti mng winding e arresto e r

Due to the presence of the interwinding capacitance and the capacitances to earth of the transformer windings, the upper elements of the transformer windings tend to be more heavily stressed than the lower portions. Due to the velocity of propagation of the impulse voltage would not be evenly distributed in the winding.

406

J R Lucas

High Voltage Testing

Due to sharp rise of the voltage of the surge, there is a large difference of voltage caused in the winding as the wave front travels up the winding. Thus there would be an overvoltage across adjacent windings. Depending on the termination, there will be reflections at the far end of the winding. If the termination is a short circuit, at the lowest point the voltage wave whose amplitude is same as the original wave but of opposite polarity is reflected. For a line which is open circuited, the reflected wave would be of the same magnitude and of the same sign. Arising out of the reflections at the far end , there would be some coils heavily stressed. The position of the heavily stressed coils depending on the velocity of propagation.

407

J R Lucas

High Voltage Testing

If flashover occurs at the gap (lightning arrestor) the voltage of the impulse suddenly drops to zero when flashover occurs. This can be represented by a full wave, and a negative wave starting from the time flashover occurs. The chopped wave, though it reduces the voltage of the surge to zero, will have a severe effect of the winding due to sharp drop in the voltage. Thus it is always necessary to subject the transformer during tests to chopped wave conditions. Generally the method is to apply full-waves and see whether damage has occurred and then to apply the chopped waves and to see whether damage has occurred and then to apply the chopped waves and to see whether damage has occurred. Example

How will the initial voltage distribution in the winding be effected if the wavefront has a duration of several microseconds ? . where 2 = Cg . calculate the ratio of the maximum initial voltage gradient in the winding to the average voltage gradient.The capacitance to earth of the complete winding is Cw. l = length of winding vx = V . Prove that the voltage at a point in the winding distant X from the neutral is a cosh x l . the neutral point being isolated from earth.408 J R Lucas High Voltage Testing A rectangular voltage V is impressed at the line terminal of a winding of a high voltage transformer . a cosh a Cw If Cg = 900 pF and Cw = 10 pF .

the voltage variation being instantaneous . . the charging up is instantaneous and the presence of the inductance of the winding may be neglected.409 J R Lucas High Voltage Testing dsurge eart Cg v line FxNw xh q In the case of a voltage with a vertical front incident on the transformer winding .

. w .410 J R Lucas High Voltage Testing Thus. = q Cw ∂x l ∂x 2 ∂ v Cg 1 ∴ .d x = .v .l = q ∂x l ∂x dx ∂q ∂v Cg ∴ = . v = V. ∂q ∂v Cg C .d x .l. 2 .d x .v = 0 2 ∂ x Cw l a a ∴ v = A cosh x + B sinh x l l a a at x = l.v. . so that v = A cosh x + B sinh x l l .

giving A = cosh a a a a V cosh . x V . d x l o l . l l 1 l V. so that q = Cw l . . sinh . . x ∂v l l l ∴ v = .411 J R Lucas High Voltage Testing ∂v a a at x = 0 .48 10 a a Maximum initial voltagegradient(at x = l) = V .a a averagestress = ∫ . = Cw l (A . sinh . tanh (9. 1 ≈ l  cosh a  l  ∴ maximumstress/ave rage stress = a ≈ 9.48 initially Substituting figureswe have a 2 = .48) ≈ V . q = 0 . cosh 0) = 0 ∂x l l V ∴ B = 0 . also = initially. cosh a l V  1  V = . sinh 0 + B . x . ∴ a = 9. cosh a ∂x cosh a 900 = 90 .

412 J R Lucas High Voltage Testing V (neutra 1 ξ F0 l) x/l x The voltage .

The winding then behaves similar to a transmission line with distributed inductances and shunt capacitances. . and the effect of the inductance during this period may not be neglected.413 J R Lucas High Voltage Testing distribution along the winding and the stress distribution (initially) are as shown in figure 6. The effect of the surge would cause a lesser stress than in the case of a surge with a vertical front. If the wave-front time is of several micro-seconds duration.13. The differential equation governing the variation of the voltage would be a fourth order partial differential equation. the charging up would not be instantaneous.

[If puncture occurs. air density etc.414 J R Lucas High Voltage Testing 6. 6.) is applied. (a) Withstand Test: The insulator should be mounted so as to simulate practical conditions. A 1/50 μs wave of the specified voltage (corrected for humidity.. (b) Flash-over test: . Flashover or puncture should not occur. The test is repeated five times for each polarity.1 Type Tests These tests are done to determine whether the particular design is suitable for the purpose.5. the insulator is permanently damaged].5Tests on Insulators The tests on insulators can be divided into three groups. sample tests and the routing tests. These are the type tests.

There shall be no damage to the insulator. This is repeated ten times. clean and dry. There should be no puncture of insulation during these tests. The voltage is gradually increased to the 50% impulse flash-over voltage. The test is done for both polarities. (c) Dry One-minute test: The insulator. shall be mounted as specified and the prescribed voltage (corrected for ambient conditions) should be gradually brought up (at power frequency) and maintained for one minute.415 J R Lucas High Voltage Testing A 1/50 μs wave is applied. Dry flash-over test: The voltage shall then be increased gradually until flash-over occurs. (d) One-minute Rain test: . There shall not be puncture or flash-over during the test.

416 J R Lucas High Voltage Testing The insulator is sprayed throughout the test with artificial rain drawn from a source of supply at a temperature within 10o C of the ambient temperature of the neighbourhood of the insulator. (e) Visible discharge test: . This is repeated ten times. There shall be no damage to the insulator. Wet flash-over test: The voltage shall then be increased gradually until flash-over occurs. The rain is sprayed at an angle of 45 o on the insulator at the prescribed rate of 3 mm/minute. The resistivity of the water should be 100 ohm-m ± 10%. The prescribed voltage is maintained for one minute.

417 J R Lucas High Voltage Testing This states that after the room has been darkened and the specified test voltage applied.. up to and including the point of breakdown.). 6.2 Sample Tests The sample is tested fully.5... (b) Mechanical loading test: The insulator shall be mechanically loaded up to the point of failure. after five minutes. (a) Temperature cycle test: The complete test shall consist of five transfers (hot-cold-hot-. This is done only on a few samples of the insulator. the load should not be less than 2000 lbf. (c) Electro-mechanical test: . there should be no visible signs of corona. When failure occurs. each transfer not exceeding 30 s.

418 J R Lucas High Voltage Testing The insulator is simultaneously subjected to electrical and mechanical stress. (e) Porosity test: . to prevent external flashover occurring. it shall be subjected to a power frequency voltage and a tensile force simultaneously. The specified overvoltage must be reached without puncture. There should be no damage caused.e. The voltage shall be 75% of dry flash-over voltage of the unit. The voltage is then gradually increased until puncture occurs. (i. (d) Overvoltage test: The insulator shall be completely immersed in an insulating medium (oil).

3 Routine Tests These are to be applied to all insulators and shall be commenced at a low voltage and shall be increased rapidly until flash-over occurs every few seconds. . for five minutes after the last punctured piece has been removed.5.i.s.419 J R Lucas High Voltage Testing Freshly broken pieces of porcelain shall show no dye penetration after having been immersed for 24 hours in an alcoholic mixture of fushing at a pressure of 2000 p. or if failures occur. 6. The voltage shall be maintained at this value for a minimum of five minutes. At the conclusion of the test the voltage shall be reduced to about onethird of the test voltage before switching off.

.420 J R Lucas High Voltage Testing Mechanical Routine Test: A mechanical load of 20% in excess of the maximum working load of the insulator is applied after suspending the insulator for one minute. There should be no mechanical failure of the insulator.

System Voltage 11 kV 33 kV 66 kV 132 kV 275 kV I. Impulse Withstand Voltage 75 kV 170 kV 325 kV 550 kV 1050 kV (1) Apply full wave impulse at 75% I.C.6Tests on Transformers The following sequence of tests is generally adopted for transformers. there should be no damage to the transformer.421 J R Lucas High Voltage Testing 6.E.E.voltage. withstand value.C.C. .E. Since the transformer should be able to withstand the I.

withstand value and observe whether there is any breakdown.E. (3) Chopped wave test at 115% fullwave amplitude : For this kind of test .E. from the waveform it is not possible to say whether any damage has taken place.422 J R Lucas High Voltage Testing The values of R and C in the impulse generator are adjusted after deriving to get the required waveform. (2) Apply full wave at 100% I. voltage and see whether there is any distortion in the waveform indicating damage.C. Since there is no voltage across insulator after chopping takes place.C. The waveform observed should be identical to applied waveform (other than for amplitude) : then the device has passed the test.(same as test 2) . the impulse generator would have to be fitted with a rod gap or controlled trigatron type gap. (4) Therefore apply full wave test again and compare the wave and at 100% of I.

423 J R Lucas High Voltage Testing (4) (4 (3 (2 (1 due 115 100 7 t 5 to partial vFigure 6.14 * ) discharge % – Test Waveforms for transformers .

424 J R Lucas High Voltage Testing .

425 J R Lucas High Voltage Testing t 0 150% 100% F .

In the American industry. there is some controversy on the requirement of this test.426 J R Lucas High Voltage Testing Since the chopped wave test exerts considerable stress on the winding. . Thus the chopped wave requirement is not universal. In this case the stress might in fact be very much more than in the British method. the chopped wave is conducted at 150% full wave and such that the chopping is done at less than the peak value.

7Tests on Cables For cables not in the supper voltage class the tests to be carried out are laid down in the appropriate British standard specifications.427 J R Lucas High Voltage Testing 6. (1) Acceptance Tests at Works (a) Conductor resistance (b) Voltage test: The applied voltage must be of approximately sinusoidal shape and of any frequency between 25 and 100 Hz. BS 480 Part I: 1954. Thus for paper impregnated insulated cables with lead or alloy sheaths. . the tests (purely electrical) are as follows. It must be increased gradually to the full value and maintain continuously for 15 minutes between conductors and between each conductors and sheath.

428 J R Lucas High Voltage Testing The required values of the test voltages are tabulated in the specification and. (ii) Between any conductor and sheath (1) Cable as manufactured. S. . Voltage Designation (1) 11 kV where Belted Cables (i) (2) (1) (ii) (2) Single-core. as one illustration of the magnitude relative to the normal voltages. & Screened Cables (ii) (1) (2) 15 kV 22 kV 24 kV 36 kV 14 kV 21 kV (i) Between conductors. the figures for the 11 kV cables for earthed system are given in the table. (2) After bending test It will be seen that a voltage test is made before and after a bending test.L.

5 kV .429 J R Lucas High Voltage Testing In this the cable has to be bent around a cylinder of specified diameter to make one complete turn: it is then unwound and the process repeated in the opposite direction .L. . 19 kV .0006 for 3–core screened cable or 0. Ionization – ie. (c) Dielectric power–factor / Voltage test (for 33 kV cables only) : Each core of every drum of completed cable is tested for dielectric power factor at room temperature at the following a.type cable. difference in power–factor between half the normal working voltage and twice the normal working voltage – shall not exceed the value declared by the manufacture and shall no case exceed 0. 38.0 kV. The cycle of process has to be carried out three times. 28.c. single phase 50 Hz voltages : 9.001 for single core and screened S. The measured power–factor at normal working voltage shall not the value declared by the manufacturer and shall in no case exceed 0.01 .5 kV.

14 kV and 11 kV for belted cable as manufactured and the value 15 kV for single core.01 at series of temperature ranging from 15o C to 65o C. S. and screened cables.430 J R Lucas High Voltage Testing The manufacturer can also be asked to produced evidence to show that the power factor at normal working voltage does not exceed 0.1 Tests on Pressurised Cables . (2) Sample test at works These include bending test above and a dripping or drainage test for cables which have to be installed vertically. Thus the value of 24 kV. 6.5 kV and 12 kV respectively.7.L. 11. become 20 kV. (3) Test when installed A voltage test similar to the above is carried out in the same manner but with some what reduced voltages.

except in the case of changes in the design. with the cable energised to 1. comprising the cable and each type of accessory to be subjected to 20 load cycles to a minimum conductor temperature 5 oC in excess of the design value. The cable to be tested at a stipulated minimum internal pressure. are stipulated for each design of cable and accessory. The dielectric thermal resistance test included in the schedule is applied only to the minimum conductor sizes.5 times the working voltage. no further type tests are required. The tests are as follows: (a) Loading cycle test: A test loop.431 J R Lucas High Voltage Testing Type approval tests. (b) Thermal stability test (132 kV cables only): . and if successful. These tests are carried out on the maximum and the minimum conductor sizes for each design and voltage rating.

comprising cable and each type of accessory to be subjected to 10 positive and 10 negative impulses at test voltage. Peak working voltage ratio during impulse test 6. to prove that the cable is thermally stable.33 times the working voltage is proposed.5 times working voltage and the loading current adjusted to give a maximum temperature 5 oC in excess of the design value. (c) Impulse test: A test loop.432 J R Lucas High Voltage Testing After test (a). Impulse test voltage 640 kV. The current to be maintained at this value for a period of 6 hours. with other test conditions unaltered. [Ex: Working voltage 132 kV. 1. the cable to be energised to 1.0] (d) Cold power-factor/voltage test: . For 275 kV cables.

(g) Binding test: The cable to be subjected to three binding cycles round a drum of diameter 20 times the diameter of the pressure retaining sheath. 6. 1. (e) Dielectric thermal resistance test: The thermal resistance of the cable is measured.8Tests on High Voltage Bushings .5 and 2 times the working voltage with the cable at the stipulated minimum internal pressure. The values not to exceed the makers' guaranteed values.5.433 J R Lucas High Voltage Testing The power factor of a 100 m length of cable to be measured at 0. The sample then to withstand the routine voltage test carried out on all production lengths of cable. (f) Mechanical Test of metallic reinforcement: A sample of cable to withstand twice the maximum specified internal pressure for a period of seven days. 1.0.

(i) Solid Bushing: A bushing consisting of a single piece of solid insulating material which is continuous between its outer surface and the inner conducting surface.434 J R Lucas High Voltage Testing 6. it includes the means of attachment to the partition. or through a ring type current transformer and insulating it there from. the oil providing the major radial insulation. The space is occupied by air. through a partition. (ii) Plain Bushing: A bushing consisting of a single piece of solid insulating material. which may be the main conductor or a conducting layer connected thereto. such as a wall or tank cover.8. . oil or other insulating medium which forms part of the insulation.1 Bushing A single or composite structure carrying a conductor or providing passage for a conductor. [See item (iii)] (iii) Oil filled Bushing: A bushing consisting of an oil-filled insulating shell. with a space between the conductor and the inner surface of the solid insulation.

in which case the intervening space may be filled with oil or other insulating medium.the lengths and diameters of the cylinders being designed with the object of controlling the internal and external electric fields.2 Tests on Bushings . (including materials impregnated with oils or other impregnants).] 6.] (iv) Condenser busing: A bushing in which cylindrical conducting layers are arranged coaxially with the conductor within a solid body of insulating materials.it is recommended that the term condenser bushing with oil filling be used for this type. [Note:A conductor bushing may be provided with a weather shield.8.435 J R Lucas High Voltage Testing [Note: The conductor may be further insulated by a series of spaced concentric cylinders which may be provided with cylindrical conducting layers with the object of controlling the internal and external electric fields.

Clause 4: A bushing shall be rated in terms of the following: a) voltage(refer table 1.e. impulse withstand voltage with 1/50 μs full wave. one minute dry withstand voltage. .e.436 J R Lucas High Voltage Testing Rating of bushings: Some of the relevant clauses from the standard is given in the following sections. For non–impulse tested bushings the rated insulation level is expressed as a power frequency voltage value i. clause 5) b) normal current (refer tables 2 and 3 clause 6) c) frequency (refer clause 7) d) insulation level (see clause 8 below) Clause 8: The insulation level of bushing is designed by a voltage which the bushing must be capable of withstanding under the specified test conditions For impulse tested bushings the rated insulation level is expressed as an impulse voltage value i.

437 J R Lucas High Voltage Testing Type Tests Clause 14: Clause 15: Clause 17: Clause 18: Clause 19: Clause 20: Clause 21: Clause 22: Power frequency test Impulse test Momentary dry withstand test (power frequency voltage) Visible discharge test (power frequency voltage) Wet withstand test (power frequency voltage) Puncture withstand test (power frequency voltage) Full wave withstand test (impulse voltage) Puncture withstand test (impulse voltage) Sample Tests Clause 23: Temperature rise test .

438 J R Lucas High Voltage Testing Clause 25: Clause 26: Clause 27: Thermal stability test Temperature cycle test Porosity test Routine Tests Clause 29: One minute dry withstand test (power frequency voltage) Clause 30: Oil lightness test Clause 31: Power factor voltage test 6.9 Tests on Porcelain and toughened glass insulators for overhead power lines .

3 kV and upwards) Classification of tests Tests are divided into three groups.439 J R Lucas High Voltage Testing Specifications B.S. pin or line post insulator which depend on shape and size of the insulator and of its metal parts and accessories. Clause 18: Impulse withstand voltage tests and 50% dry impulse flashover test Clause 19: Power frequency voltage one-minute wet test and wet flashover test Clause 20: Visible discharge test .137:1960 (3. as shown. Tests in Group I (Type tests) These tests are intended to verify those characteristics of an insulator or set. They are normally made once only to establish design characteristics.

Clause 23: Verification of dimensions Clause 24: Temperature cycle test Clause 25: Mechanical failing load test or Clause 26: Electro-mechanical failing load test Clause 27: Overvoltage tests Clause 28: Porosity test on porcelain insulators . They are made on insulators taken at random from every batch offered for acceptance.440 J R Lucas High Voltage Testing Tests in Group II (Sample tests) These tests are for the purpose of verifying certain characteristics of a string insulator unit.line post insulator or pin insulator and pin and the quality of the materials used.

Tests in Group III (Sample tests) These tests are for the purpose of eliminating insulators with manufacturing defects.S.729 and shall satisfy the requirements of that standard. Clause 30: Galvanizing test: The galvanized samples shall be tested in accordance with B.441 J R Lucas High Voltage Testing Clause 29: Thermal shock test on toughened glass insulators: The glass shall not shatter when the sample insulators are completely immerses in water at a temperature not exceeding 50 C. the temperature of the insulators immediately before immersion being at least 100 C higher than that of the water. They are made on every insulator offered for acceptance. .

. for a period of not less than 10 seconds.442 J R Lucas High Voltage Testing Clause 32: Electrical test on porcelain insulators Clause 33: Thermal shock test on toughened glass insulators Clause 34: Mechanical test on string insulator units: Every string unit shall be subjected to a tensile load of at least 40% of the specified minimum failing load.

Sign up to vote on this title
UsefulNot useful