Modern
ART
explored
...p3
Tackling
TDR
theory
...p4 - 5
Power
problems?
Try mice!
p8
Published by Megger
April 2014
ELECTRICAL
TESTER
Moisture measurement
The industry’s recognised information tool
analysis will produce an exaggerated error in
the moisture-in-insulation result.
else. Note that, without the temperature, no
calculations can be done to determine the
moisture in paper or relative saturation. The
only information you will have is that the oil
has moisture in it!
Tom Dalton - Business unit manager:
power transformers, Martec, South Africa
Introduction
It is well documented that moisture in a
mineral oil cooled and insulated power
transformer has detrimental effects. In fact,
it is said that doubling the moisture content
in the transformer will have the effect of
approximately halving the life of the unit.
Thus the transformer will deliver only half the
expected return on investment and its reliability
will be impacted earlier in its life. This article
looks at some of the methods of determining
the amount of moisture that will affect the
operation of the unit and the subsequent
management of the oil and paper systems.
Moisture effects in an operating
transformer
When a transformer is delivered to a client the
insulation should be dried to around 0.5% (by dry
weight). During the operation of a transformer
there are a number of factors that will influence
the gradual production and contamination of
the system. Moisture has a profound effect on
mineral oil, and will cause the dielectric strength
of the fluid to drop considerably.
First, atmospheric moisture will have an impact,
especially in units that are open breathing
(breathe to atmosphere via a breather filled
with a desiccant). If the desiccant is not
maintained correctly, the oil will absorb
moisture from the atmosphere as the
transformer temperature cycles. The hotter the
oil, the more moisture it will absorb.
Second, whether transformers are open
breathers or sealed units (sealed meaning
that the unit does not breathe to atmosphere
directly), paper degradation by-products and
natural gassing of transformer oil will produce
moisture in small quantities (approximately
0.5 – 1 ppm per year). Note that transformers
that do not use paper-based insulating material
are less susceptible to this phenomenon, but
the natural gassing of the oil will still produce
small amounts of moisture. Coming back to
the breakdown by-products of paper, an OH
molecule is given off when the cellulose chain
www.megger.com
is severed by heat and electrical stress. With
most insulating fluids some hydrogen is given
off during normal operation and even greater
quantities are produced under overload and
fault conditions.
From an insulating fluid perspective: as most
insulating fluids are hydrocarbon based, small
amounts of hydrogen are given off during
operation. This combines with the ever-present
oxygen (a natural marriage) to form H2O –
moisture.
With more moisture in the system, temperature
cycling brings further destruction and thus
the deterioration increases over time. Without
intervention, the life of the transformer will be
severely impacted.
Monitoring
As a transformer operates, moisture will
move from the insulation body (thin and thick
insulation) into the oil as it heats and will move
back to the insulation from the oil as it cools.
This phenomenon is called equilibrium. If the
transformer loading and ambient temperature
were to remain constant for a long period of
time, eventually moisture movement would
cease and a state of equilibrium would be
reached. The insulation system will always
seek to obtain an equilibrium state, but
with constant load changes and ambient
temperature fluctuations, this hardly ever
occurs in real life. This is also very difficult
to predict as different parts of the system
(both oil and solid insulation) are at different
temperatures. Thus equilibrium is merely an
assumption of where the moisture may be at a
specific point in time or temperature.
To add to the dilemma, thin and thick insulation
will give off moisture at different rates (diffusion
rates). With thin insulation (mainly paper)
the transfer is quick, but with thick insulation
(blocking) it is much slower (diffusion rate is
higher for thick insulation). Note that in Figure
1 the set of curves depicting the equilibrium
and moisture content in insulation assumes
that all insulation is the same in physical size
and dimensions. Also note that at lower oil
temperatures, the chart becomes inaccurate.
Any small change in moisture-in-oil will produce
large changes in moisture content of the
insulation, thus any small inaccuracy in the
There are a number of methods used to
monitor moisture in transformers. Traditionally
an oil sample would be extracted from the
transformer and sent to a laboratory for
analysis. However, there are potentially flaws
in the process, which mean that the results
are not always reliable. These flaws are
introduced early in the process, typically at the
sampling stage, which can introduce moisture
and contaminate the sample. Transporting
the sample in a tin also poses the risk of
atmospheric contamination due to temperature
cycling. The accuracy of the Karl Fischer
testing at the laboratory also plays a role in the
accuracy of the result received. However, there
are techniques that can be used to reduce
these flaws and a good point to start with is
training the person who takes the sample.
Furthermore, care should be taken when
choosing the containers in which moisture
samples are taken. Traditionally, square or
round tins are used but these are not the best
choices – glass syringes are better for sampling.
Then there is the question of what should
be done with the result when it is received.
Moisture in oil is very dependent on the
operating temperature of the transformer,
especially when the sample was taken. Without
the transformer’s temperature reading at the
time of sampling, the result obtained only
relates to the moisture in the oil and nothing
Figure 1: Typical moisture equilibrium curves
The foregoing discussion explains why moisture
assessment using an oil sample alone is not
always adequate. An understanding of the
issues discussed, however, will help in choosing
a better methodology to measure moisture
wherever it may be.
Moisture in oil measurement
As stated in the previous section, traditionally
an oil sample is taken and analysed to obtain
a result. Don’t throw this data away as it is
still useful! With correct sampling techniques
one can obtain good results. However, it
is suggested that this data is used as the
first line of monitoring, which, with some
applied thought, can paint a useful picture
in understanding the moisture status and
trigger further action. In Figure 2 it can be
seen that plotting the data on a set of inverted
equilibrium curves is useful. If most of the
plotted points fall below the green line, the
transformer is generally dry. If, however, the
data points fall between the two curves, it is
a warning to take further measures. Lastly, if
the majority of data points fall above the two
curves, the insulation and oil are generally wet
and corrective action is necessary.
If the data points are in the category between
and/or above the two curves, an alternative
method of moisture-in-oil measurement can be
employed. This method is an excellent option as
it can be done on-line and has benefit in that it
produces real time data.
By employing a moisture-in-oil probe to
measure the dynamics of the moisture and
temperature at the same time, it is easier to
detect when too much moisture is leaving the
solid insulation system, and the rate at which it
is being given off by the solid insulation. Using
this method, the moisture movement can be
tracked and monitored over a period of time (a
week is preferable). In conjunction with moisture
measurement, it can be determined what the
exchange rate is when the load is fluctuating.
Monitoring it for a week will reveal loading
patterns in most cases. These patterns will usually
repeat the cycle every week and show up any
sharp increases of moisture movement.
continued on page 2
Figure 2: Moisture data plotted on a set of inverted equilibrium curves (2.5% - 3%)
ELECTRICAL TESTER - April 2014
1
ELECTRICAL
TESTER
The industry’s recognised information tool
Contents
Moisture measurement.......................... P1-3
Tom Dalton - Business unit manager: power
transformers, Martec, South Africa
ART - Attached Rod Techniques .............P3
Paul Swinerd - Product portfolio
manager - Power
Time Domain Reflectometers
- the physical basics....................................P4
Peter Herpertz - Product manager - power,
SebaKMT
TDR fault finding: cable fault basics....P5
Peter Dennis - Product manager communications
When reflection isn’t the answer............P6
Peter Herpertz - Product manager,
power, SebaKMT
University at Buffalo
collaboration brings benefits for all.....P6
Casey Henry, marketing program manager
Putting cables to the test...... ...................P7
Clive Pink - Product manager
Timely testing helps restore
power to 200,000 .......................................P7
Erik Blichfeld - Produktchef, SebaKMT A/S
Q&A...................................................................P8
The stability factor......................................P8
Damon Mount - power sales manager
Power problems? Bring on the mice!....P8
Ted Kim, Regional sales manager, North Asia
continued from page 1
Note that the probe’s location and oil flow are
extremely important. The probe must be placed
in a location that has rapid oil flow or at least
a steady flow over the probe tip. Normally it
would be placed in the cooling system (inlet to
or outlet from the cooler bank) or in pumped oil
flow. Another location is in the flow of the online gas-in-oil monitor. The bottom main tank
sampling point is not always the best location
as there is little movement over the tip of the
probe and in this case the measurement would
merely measure the oil close to the probe tip.
One of the outstanding benefits of this method
is the rate of change. This is an important
factor and especially important if rapid changes
in load and/or temperature are occurring. Too
much moisture in the system with rapid load
changes can cause detrimental conditions
leading to disastrous results. With rapid load
growth and wet insulation, there is a dynamic
that leads to insulation failure very quickly.
The transformer is cold and the oil is cold, with
the moisture predominantly embedded in the
solid insulation. Sudden high loading will drive
moisture out of the solid insulation rapidly and
the oil, not being able to absorb the now free
moisture, will have a low dielectric strength
zone where the moisture cannot be moved
away (high saturation zone). A characteristic of
oil is that at low temperature it is not capable
of absorbing the quantity of moisture being
driven out of the paper and will only be able
to do so once a higher temperature is reached.
In an operating transformer, the volume of oil
takes time to reach higher temperature. (Like
a kettle put on to boil, there will be aggressive
heating of the moisture near the element but
water has not yet boiled – it takes time.) This
creates a very low dielectric strength in areas
where there is insufficient oil flow (see Figure
3). To add to the problem, when oil is cold
the viscosity is higher (thicker) and the oil is
then sluggish and does not flush the moisture
away. This set of conditions can often lead to
insulation system break down and flashover.
When you have finished with
this magazine please recycle it.
The rights of the individuals attributed in Electrical Tester to
be identified as authors of their respective articles has been
asserted by them in accordance with the Copyright, Designs
and Patents Act 1988. © Copyright Megger. All rights reserved.
No part of Electrical Tester may be reproduced in a retrieval
system, or transmitted in any form or by any means, electronic,
mechanical, photo-copying, recording or otherwise without the
Figure 3: Moisture in highly loaded transformers
prior written permission of Megger.
To request a licence to use an article in Electrical Tester, please
email electricaltester@Megger.com, with a brief outline of the
reasons for your request.
All trademarks used herein are the property of their respective
owners. The use of any trademark in this text does not imply
trademark ownership rights in such trademarks, nor does use
of such trademarks imply any affiliation with or endorsement of
Electrical Tester by such owners.
A printed newsletter is not as interactive as its email equivalent
so to help you find items quickly on www.megger.com, we have
underlined key search words in blue.
Note from the Editor - Time for your say.
We have introduced a ‘Questions and Answers’ section and
would like your input. If you have any questions or stories that
you think we could use, then please email
electricaltester@megger.com
‘Views expressed in Electrical Tester are not necessarily
the views of Megger.’
The word ‘Megger’ is a registered trademark
Editor Nick Hilditch.
T +44 (0)1304 502232
E nick.hilditch@megger.com
Megger Limited
Archcliffe Road Dover Kent CT17 9EN
T +44 (0)1304 502100
E electricaltester@megger.com
www.megger.com
2
Most on-line dissolved gas analysers have
moisture detection built in and will measure
moisture along with the gases, however it is
important that a temperature-in-oil probe is fitted
and the temperature monitored along with the
moisture. From this data a very good idea of the
relative saturation can be calculated and this
parameter will be most useful in determining the
state of the transformer in terms of moisture.
Using moisture in oil measurement to determine
moisture in cellulose is a tricky business as the
equilibrium plays a major role and because
different parts of the transformer are at
different temperatures and states of equilibrium.
In most cases a set of equilibrium curves are
used to determine the amount of moisture in
paper. There are a few technical papers that
attempt to evaluate the mechanism, but there is
still much doubt as to their accuracy.
Moisture in paper measurement
Moisture in cellulose is a difficult parameter to
measure. There are two main methods to determine
this and the results are not always reliable.
First, taking a paper sample means that the
unit must be out of service. Either a hatch on
the transformer has to be opened and some
paper (or board) removed from the unit, or
the unit must be transferred to a workshop
environment where it is easier to access
paper and board. However, in both cases it
is necessary to contend with atmospheric
ELECTRICAL TESTER - April 2014
lumped together by connecting all the HV
terminals (all phases), together, all the LV
terminals (all phases) together and all the
tertiary terminals (all phases) together, to
produce three entities. Typically three tests
conditions that will influence the outcome
of the analysis. It takes a skilled technician to
perform this type of test, and he/she needs to
ensure that the location where the insulation
was removed is restored. Furthermore, the
sample has to be handled extremely carefully.
Any outside influences, such as atmospheric
conditions and poor handling will contaminate
the sample and render it useless, giving
incorrect results. Temperature and relative
humidity at the time of taking the sample will
have a significant impact.
The insulation’s diffusion rate plays a key role
in the transfer of moisture between insulation
and oil. Larger blocking or thick insulation is
not as badly affected and good results can be
obtained if it is handled correctly. Larger blocks
of insulation are normally affected by surface
moisture, e.g. between 1 – 1.5 mm deep.
Deeper moisture is locked in and will take
much effort to release, ie longer periods of
high temperature and vacuum.
Using this method of moisture determination
outside of a controlled environment is
challenging at the best of times. There are too
many obstacles to make this a cost-effective
way of measuring moisture.
Figure 4: Data obtained from a FDS test showing
how the different parameters affect the curve
– CHL, CL and CH – are performed on the
transformer. The CHL measurement is preferred
for moisture and oil assessment.
Moisture in air measurement
In this method of measurement there are
again some questions as to the accuracy of
the results. This method can only be used to
determine the moisture in air, and to some
degree moisture in insulation. Here again, there
is the question of equilibrium state and, as
mentioned above, there is still some doubt as
to when all the components in a transformer
are in equilibrium.
To perform this measurement the transformer
must have all the oil drained out, so the
measurement cannot be performed online.
This technique is normally used during
manufacture and repair either in a workshop
or on site. The unit is filled with extremely
dry air and left to stand for at least 48 hours
with as little temperature variation as possible.
It is important to ask if the insulation is “oil
impregnated” or “dry” as these parameters
will have a significant effect on the results. Dry
paper has the ability to transfer moisture far
quicker than oil impregnated paper.
A dew point probe of high accuracy is needed
and it must be installed in such a way that it
has air flowing over the tip. The transformer
tank must be pressurised with dry air (dew
point temperature below -50° C) to greater
than atmospheric air pressure (typically 25
kpa) and left to stand for a least 48 hours.
The duration of the measurement should
be between 10 – 15 minutes and the data
logged over that time period. Note that if the
technician performing the test touches the
probe with his/her fingers the initial reading
will be significantly higher until that moisture
has dissipated. If this happens the test duration
should be lengthened. Once the data has been
captured, the initial data should be discarded
and ideally the flatter section of the data taken
and averaged. The averaged data must then be
applied to the “Pipers” chart and the moisture
in paper read off the chart.
Figure 5: Typical measurement connection
Depending on the instrument, the test will
vary the frequency between 1000 Hz down to
about 0.001 Hz (the preferred frequency range
depends on the insulation temperature and
can be set by the test technician). Once the
tests are completed the data set measured is
modelled against a predefined set of curves and
the closest curve is matched to the data. From
the modelling performed, the instrument will
give an accurate determination of not only the
moisture in insulation but also moisture in oil ie
system moisture. Figure 6 shows typical curves
measured and the various observations made.
Figure 6: Typical traces for varying conditions
FDS can be used in dry air (no oil in the
transformer tank) and with oil filled
transformers, making it a very useful tool in
both field and workshop applications.
This method gives an indication of surface
moisture, but there is still doubt about its accuracy
for determining the total amount of moisture in
the transformer insulation and there are better
means of determining moisture in paper.
System approach
The system approach is far more refined
and uses an electrical means to measure
the “system” rather than trying to measure
one of the components (either oil or air)
and calculating the resultant moisture. This
method is finding greater acceptance and is
improving as the technology matures and gains
momentum. Frequency Domain Spectroscopy
(FDS) or variable frequency dissipation factor
measurement (tan delta) takes both oil and solid
insulation into consideration. The instrument
uses the data measured (dielectric dissipation
factor or tan Δ) and models this to a known
curve, which then equates to moisture in oil and
moisture in insulation. That is, it takes the whole
insulating system into consideration.
This measurement cannot be done on line and
the transformer will need to be disconnected
from the network. The windings are normally
Figure 7: Typical moisture analysis performed
with modelling
Management
Many people ask how much moisture should
be allowed or is good practice. To answer this
question three categories of transformer need
to be considered:
„„New
„„In repair process
„„In service
www.megger.com
ELECTRICAL
TESTER
The industry’s recognised information tool
A new transformer should be dried to a value of
0.5% or below.
A point to remember is that there is a tradeoff, because to dry the unit to very low values
takes a number of cycles to dry the windings.
Depending on the methods and technique,
these drying cycles have a tendency to lower
the degree of polymerisation (the shortening
of fibres in the insulation material) and thus
shorten the life of the unit’s insulation. Also
some manufacturers oil impregnate the
windings early in the process and thus it is more
difficult to remove moisture in the drying cycle,
but this works both ways as the insulation will
not absorb moisture as readily. However, left
un-impregnated the insulation will get wetter
quicker and it will take longer to remove the
moisture. Careful monitoring of air condition in
the manufacturing plant is necessary to prevent
the cellulose-based windings from absorbing
moisture. With the FDS technique, both oilfilled and non-filled units can now be tested
after tanking.
It is suggested that FDS measurements are
made prior to testing the transformer in the
manufacturing plant, on arrival at site, and prior
to energising.
Transformers that have been repaired should
be dried to between 0.8 – 1.2%, but these
values will depend on the wetness of the unit
prior to being repaired. Note that the degree
of polymerisation plays a major role in this
decision. If the degree of polymerisation is
found to be low, the dry-out may damage the
insulation to the point of no return. Multiple
drying cycles will deplete the paper life, since
the more heat applied, the greater will be the
ageing effect on the cellulose. If the cellulose’s
degree of polymerisation is already low then
drying to an unnecessarily low value will only
cause further loss of life!
It is suggested that the unit should be tested
prior to and after repairs, and also prior to
energising.
Transformers in service should typically be < 2%
for large units (e.g. 100 MVA and above) and <
2.5% for smaller units.
Good maintenance practice is to test the
transformer every 2 to 5 years with the FDS
method but as stated, retain the moisture-in-oil
data to keep your finger on the pulse.
In summary
Keeping transformers dry is the preferred
practice. To do this, a molecular sieve or other
on-line drying technology can be deployed for
continuous drying of the oil and solid insulation
system, thus avoiding the need for major dry
outs when the unit is found to be wet.
Measuring moisture in an operating transformer
is not practical without the oil temperature
being taken along with the sample. However,
for reliable results it is best to have a trained
sampler to take the sample from the
transformer using the correct techniques and
equipment, and transporting the sample reliably
to the laboratory. Oil sampling is still a good
first line defence, but follow up measurements
must be made if the transformer shows signs of
undue moisture increases that are unrelated to
variations resulting from the sample process.
The other methods that are mentioned in this
article are second tier methods and are used
to gather further and more detailed condition
information. However, in many cases these
methods have their own limitations. A good
solution is the frequency domain spectroscopy
technique which has made this uncertainty
a thing of the past. Spectroscopy and data
modelling offer clear advantages in modern
science and laboratory practice, and now FDS
has brought this modern laboratory technology
into the field, complemented by well-proven
modelling techniques that provide very specific
decision support for the user.
Different asset owners have different priorities
for the operation and maintenance of
transformers. Some, for example, focus on
optimisation and return on investment rather
than reliability, whereas for others reliability
is paramount. In reality, however, a balanced
approach is best.
In closing, it is always worth remembering that
a wet transformer is an unreliable transformer.
www.megger.com
ART - Attached Rod Technique
in earth / ground testing
Figure 3: Example of Earth coupling
as a result of overlapping spheres of influence
to the building. As a result, the sphere of
influence of the electrode and that of the
building overlap. In effect, “earth coupling” is
occurring and this changes the
equivalent circuit.
The overlapping of the spheres of influence
introduces additional impedances that make
resolving the resistance of the electrode difficult
using ART. The most likely result will be either a
“clamp low” warning or an unexpectedly high
electrode resistance reading. In cases like this,
the traditional three-pole test method should be
used, with the electrode under test disconnected.
Figure 1: ART - principal of the technique
Paul Swinerd - Product portfolio
manager - Power
The art of earth electrode testing
The testing of earth systems has, for many years,
principally relied on the tried-and-tested fall-ofpotential method and similar techniques. These
techniques give reliable results, but they can
be time consuming. To measure the resistance
of an individual earth electrode, it is necessary
to disconnect that electrode from the rest of
the earth system. Not only does this take time,
disconnecting the electrode may also compromise
the safety of the installation it is protecting.
To address these problems, a novel form of
earth electrode testing has been developed.
This is known as the Attached Rod Technique,
or simply ART.
When an earth tester injects a test current
into an earth electrode that is still connected
to the earthing system, the current flows not
only into the electrode under test but also into
other electrodes that are connected in parallel
and into any other available paths to earth.
However instruments with ART capabilities use
a current clamp (iClamp) to directly measure
the current flowing in the electrode under
test. The instrument then uses this current
to calculate the electrode’s resistance. No
disconnection is needed, so there’s no wasted
time, no unnecessary plant downtime, no
inconvenience and no scraped knuckles!
the electrode under test can have a resistance
of up to 20 times that of the total system and
ART will still give reliable measurements. If the
resistance is more than 20 times the resistance
of the total system, the traditional three-pole
measuring method should be used.
Do not, however, underestimate the usefulness
of ART! If the current in the electrode under
test is less than 5% of the total test current,
the tester will display a “clamp low” warning.
If, under these conditions, users measure the
resistance of the complete earth system using
the three-pole method, they will know that
the resistance of the electrode under test is at
least 20 times this value. Often, this is enough
information to decide whether or not the
resistance of the electrode is satisfactory.
There is, however, another factor that needs
to be considered when using ART testing and
this relates to the spheres of influence around
the earth electrode(s) and building earth paths,
which may be through water or gas pipes, or
through the metal framework of the building.
Consider the situation shown in Figure 2. Here,
In Figure 1, the test current is injected between
points X and C. The instrument measures the
voltage between points X and P at the test
frequency only, which enables it to ignore the
effects of other currents that may be flowing
in the earth system. It then uses Ohm’s law to
calculate and display the electrode resistance.
As can be seen from the diagram, the addition
of the iClamp allows the current in an
individual electrode to be measured separately.
The iClamp also responds only to currents at
the frequency produced by the instrument,
allowing other currents flowing in the electrode
to be ignored.
In practice, ART testing works well provided
that the current in the electrode under test, as
measured by the iClamp, is at least 5% of the
total test current. To look at it in another way,
There are a few other situations where ART
testing is unsuitable, and one of these is
illustrated in Figure 4, which shows guy lines
connected to a metal tower. The problem may
not, at first, be obvious.
An attempt is being made to measure the earth
resistance of the guy line that has the iClamp
attached, but all of the guy lines on this tower
are shorted together. This means that the
current being measured by the iClamp is not
only flowing to ground at the anchor point, but
is also flowing back up the other guy lines and
then to earth via the tower. ART testing will,
therefore, give an incorrect result.
To avoid problems of this type, always consider
carefully where the test current will flow. For
successful ART measurements, the current from
the electrode under test must flow only into the
soil mass surrounding the electrode.
Let’s look a little more closely at how a typical
earth electrode resistance tester with ART
functionality works. Figure 1 shows
the essentials.
Testers with ART functionality are also capable
of traditional three-pole fall-of-potential
measurements. When used in this way they
inject a test current at a frequency that has
been chosen so that it doesn’t clash with the
power frequency or its harmonics. A frequency
of 128 Hz, which avoids the harmonics of both
50 Hz and 60 Hz supplies, is often used.
(More detailed information about spheres
of influence can be found in the publication
“Getting Down to Earth”, which is available as
a free download www.megger.com. Simply log
in or register, and navigate to the publications
section.)
Figure 2: Spheres of influence
the spheres of influence are separate from each
other, and the equivalent circuit is, therefore,
as shown at the bottom of the figure. In cases
like this, ART testing will work well. Subject, of
course, to the 20:1 rule.
This article has outlined how ART earth
electrode testing works and has discussed some
of its limitations. It is important to remember,
however, that there are very many applications
where ART testing works extremely well.
Examples include earth farms, pole-mounted
transformers, domestic TT installations, single
guy lines on towers and lightning protection
electrodes. In short, provided its limitations are
clearly understood, ART testing is an invaluable
tool that saves time, money and trouble.
Now consider the situation shown in Figure
3. Here the electrode under test is very close
Figure 4: Mixed readings can be caused by multiple guy-lines
ELECTRICAL TESTER - April 2014
3
ELECTRICAL
TESTER
The industry’s recognised information tool
Time Domain Reflectometers
- the physical basics
Figure 1: Oversampling technique
Peter Herpertz - Product manager,
power, SebaKMT
When choosing or using a time domain
reflectometer (TDR), it is very useful to have
at least some knowledge of the theory that
underpins reflectometer operation and the
technology that is used to turn that theory into
a practical instrument. The objective of this
article is to provide this essential knowledge in
a concise form that can be readily related to
real-world requirements and applications.
As a starting point, it is useful to note that
reflectometers can be divided into two main
groups – instruments that are intended for
use on power cables (power reflectometers)
and those that are intended for use on
telecommunications systems (telecom
reflectometers). The essential difference is that
power reflectometers use flash A/D conversion
technology, while telecom reflectometers use
sampling technology.
Power reflectometers
Power reflectometers have at their heart a fast
and expensive flash analogue-to-digital converter.
The latest instruments have sampling rates of up
to 400 MHz, which is at least twice the rate used
in the previous generation of instruments.
It is widely believed that the higher the
sampling rate, the better the instrument,
because higher sampling rates provide
increased resolution. In theory this is true,
but in practice there is little benefit to be had
from further increasing sampling rates, as the
resulting higher resolution is only relevant
at short distances. At larger distances, the
instrument display cannot show the full
measured resolution.
(80m/μs)80.000km/s
Res =
= 0.2m
400MHz
Range and resolution
To explore in more detail the relationship
between sampling rate, range and display
resolution, we will examine the situation with
a modern power reflectometers that is among
the best of its type – the Teleflex SX or its large
equivalent the Teleflex VX.
Assuming that the measuring pulse travels with
a typical propagation velocity of 80 m/µs, with
a sampling rate of 400 MHz the theoretical
resolution will be 0.2 m, as this equation shows:
The lowest range of the Teleflex SX is 20 m.
The display is 1,024 pixels wide so, when the
20 m range is in use, in round terms 20 m =
1,000 pixels. This means that 0.2 m = 10 pixels.
On this range, the full theoretical resolution
of the instrument can, therefore, be used.
However, when the 200 m range is selected,
0.2 m is equivalent to only one pixel, so the
usable resolution is starting to be limited by the
display rather than the sampling rate. At longer
ranges, even if the display is magnified with the
zoom function, the limiting effect of the display
is completely dominant and the theoretical
resolution of the instrument calculated from
the sampling rate alone can never be used.
Impulse covering zone
Another factor that affects measurement
resolution is the pulse width. With the smallest
4
Figure 2: Full sampling
pulse width of 20 ns and a propagation
velocity of 80 m/µs, the pulse width is
equivalent to 1,6 m. Similarly, a pulse width of
10 µs is equivalent to 800 m. These distance
equivalents are known as the covering zone
(CZ) for a particular pulse width and are easily
calculated as:
lmpcz = impulse width [µs] * propagation
velocity [m/µs]
The pulse width is typically related directly to
the range being used on the instrument.
Some leeway – one step up or down – is
possible, but in general a pulse that is too
short will be lost because of cable attenuation,
while too large a pulse will limit the useful
resolution. Taken together the resolution and
the pulse width determine the accuracy of the
measurement made by the instrument, which
is typically around 0.1%.
This means, for example, that for a given pulse
width, events with a shorter distance between
them than the covering zone cannot be
resolved. They are, therefore, either not visible
at all or visible only as merged reflection of
signals from both events.
Principles of power reflectometer
flash conversion
A flash converter takes the whole returning
analogue signal and converts it continuously
at a high conversion rate into a digital
signal. Ideally, it should complete the whole
conversion process in a single pass, as shown
in Figure 2.
However, to reduce costs, some instrument
manufacturers use “oversampling” technology.
This typically means that the A-to-D converter
operates at one quarter of the claimed
sampling rate, but runs sequentially four times.
As is shown in Figure 1, this means that, in
reality, the instrument makes four different
measurements – those in the first pass are
shown in blue, those in the second pass in red,
those in the third pass in green and those in
the fourth pass in orange.
For example, an instrument with a 50 MHz
converter and four-times oversampling will be
presented as an instrument with a sampling
rate of 200 MHz. It is an indisputable fact,
however, that oversampling does not provide
the same quality as a single pass measurement.
With oversampling, after the four measuring
cycles, the instrument interpolates the
measuring points. Real signal details can be
lost in this interpolation process, and spurious
details can be generated.
An arcing fault, for example, changes very
rapidly and its signature can be completely
different on consecutive oversampling cycles.
This can be seen in Figure 1 where only the
blue conversion cycle has changed, producing
a completely different signature.
instant, producing one “pixel” of the trace.
This process is repeated until the complete
trace has been recorded. This method of
capturing information is much slower than
the flash conversion method used in power
reflectometers, but it yields more accurate
information. Fast events like arc reflection are,
however, unlikely to be captured.
In telecom fault location, high accuracy is
more important that high measuring speed
because pinpointing is much more difficult on
telecom cables than it is on power cables. The
construction of telecom cables does, however,
provide greater consistency for important
parameters like propagation velocity, making it
easier to determine an accurate distance to the
fault using reflectometry. This high prelocation
accuracy makes pinpointing somewhat easier.
immediately triggers the reflectometer. In almost
all circumstances, this provides a perfect trace
on cables up to several kilometres in length.
There are, however, always exceptions. These
typically relate to very long cables or cables
affected by water/humidity problems. In these
cases, there is a delay in response either due
to the extended signal travelling time on long
cables, or the unpredictable propagation
velocity of cables affected by water. Here
another technology – ARMslide – comes into
its own. This records up to 15 traces during
one ARM discharge, which ensures that at
least one trace will provide the required data.
To deal with the most challenging applications,
it is even possible to adjust the trigger timing
to intermediate values between the 15
measurements.
Triggering
With power reflectometers, a widely used
measuring technique is the arc reflection
method (ARM), where a surge generator is
used to produce an arc at the fault location,
and this arc (while burning, it is low resistive)
reflects the pulse from the reflectometer. For
this method of measurement to work well, the
reflectometer needs a very reliable trigger.
The conventional triggering method is delay
time adjustment, where the instrument user
manually adjusts the delay time between
the power surge that strikes the arc and
the release of the measuring pulse from the
reflectometer. This approach requires the user
to have a high level of skill as the delay time
needed depends on several parameters related
to the ARM oscillation, which is determined
by the resonant circuit formed by the surge
capacitor, the cable capacitance and the cable
inductance. These factors change with cable
length, so it’s clear that a fixed delay time
is not a reliable solution. The objective is to
trigger the reflectometer when the arc from
the ARM discharge has reached its highest
current value and is burning solidly. One
method of compensating for the difficulty of
triggering at exactly the right moment is to
take several reflectometer recordings of the
same ARM shot. This increases the probability
of capturing at least one useable trace.
A better approach, however, is the ∆U
triggering system, which is also known as “one
shot triggering technology”. This works by
continuously evaluating the current of the ARM
oscillation. After the current has reached its
peak, the instrument waits for the voltage to fall
by a predetermined amount (∆U) and then
Figure 3: ∆U triggering (above) and ARMslide
recording (below) showing the effect of timing
adjustments
Range compensation
For both power and telecom reflectometers,
the attenuation and dispersion of the cable
cause problems. Impedance and resistive
losses, cable length and other influences
change the amplitude and shape of the
pulse transmitted by the cable. Attenuation
makes the signal smaller as distance
increases. The effect of attenuation is
shown by the red line in Figure 4. Since
attenuation follows an exponential function,
it can be calculated and compensated
for. The distance-related amplitude
correction provided as part of the ProRange
compensation function is shown in Figure 5.
Dispersion is another factor that affects
pulse shape. Higher signal frequencies are
attenuated more than lower frequencies,
Telecom reflectometer technology
As has already been mentioned,
reflectometers designed principally for use in
telecommunications applications work in a
different way from those designed for use in
power applications. Telecom reflectometers use
sampling technology, which basically means
that information about points along the cable
is captured sequentially. A measuring pulse
is sent out into the cable, and the amplitude
of the return pulse is recorded at a specific
ELECTRICAL TESTER - April 2014
Figure 4: Cable attenuation and dispersion
www.megger.com
ELECTRICAL
TESTER
The industry’s recognised information tool
with a result that distant reflections appear
wider and lower in amplitude than nearby
reflections. The combination of attenuation
and dispersion mean that very distant
reflections are sometimes difficult to recognise
and evaluate.
TDR fault finding: cable fault basics
Figure 5: Distance related amplitude correction
The ProRange function allows all events
to be displayed with the correct amplitude
irrespective of their distance, which makes
evaluating the events much easier. Dispersion
still affects the trace, but has far less influence
on event visibility.
Dead zone
The dead zone is the basis for endless
discussions. It results from the fact that the
transmit pulse, which looks to the instrument
just the same as a reflected pulse except
that it is much larger in amplitude, saturates
the instrument for a certain distance at
the beginning of the trace. Only beyond
this distance do reflected pulses become
visible. Various technologies are available
to compensate for the dead zone, one of
the most effective being the use of a split
transformer. The basic circuit for this type of
compensation is shown in Figure 6.
Peter Dennis - Product manager communications
Introduction
A TDR (time domain reflectometer) uses the
radar principle to identify faults on cables.
The instrument fires a pulse down the cable,
and any change in the impedance of the
cable will result in reflections being sent
back down the cable toward the instrument.
These reflections are measured and displayed
so that a “map” of the cable is shown. Many
faults are found at terminations, cable joints
and other locations where there has been
disruption to the route of the cable.
Since they work by identifying changes in
impedance, TDRs require two conductors
that run parallel to each other in order
to operate. Any connection, change of
cable type, break in the cable, or fault will
cause a change of impedance. Each type
of change has a different effect on the TDR
display; a positive reflection shows higher
impedance, a negative reflection shows
lower impedance.
Using a TDR
Identify the location of fault by testing from
both ends of the cable.
Figure 6: Dead zone compensation
With this arrangement, the variable resistance
R is compared with the impedance of the
cable, and is adjusted until the impulse
currents I1 and I2 are as nearly as possible
identical, which means that they cancel each
other out in the transformer. The reflectometer
does not, therefore, see the measurement
pulse.
In practice, R should be set using the
reflectometer’s lowest measurement range
and should be adjusted until the positive and
negative reflections are equal in size and as
small as possible in amplitude. The returning
reflected pulses produce current I3 in the
transformer, and are completely unaffected by
the compensation circuit.
Even better dead zone compensation can be
achieved in telecom reflectometers, when a
second cable pair that is nominally identical
with the pair under test is connected to the
second input (Z2) to provide compensation.
Since both cable pairs have identical
parameters – at least in theory – they will
cancel each other out completely except for
differences that represent the fault on the pair
under test.
The next article in this series will look at
measuring techniques using reflectometers,
and will include a range of practical examples
showing typical traces that are produced by
faults of various types.
www.megger.com
Identify the faulty cable with an insulation
tester. If the fault is low resistance,
Cable type
Construction
Velocity factor %
Velocity factor m/μs
Power
Paper oil filled
0.72 to 0.84
216 to 252
Power
XLPE
0.54 to 0.62
162 to 186
Power
EPR
0.45 to 0.57
135 to 171
Twisted pair
Polyethylene
0.64 to 0.67
192 to 201
Twisted pair
PTFE
0.71
213
Twisted pair
Paper
0.72 to 0.88
216 to 264
Telecomms
PIC
0.65 to 0.72
195 to 216
Telecomms
Pulp
0.66 to 0.71
198 to 213
Telecomms
Gel filled
0.58 to 0.68
174 to 204
Telecomms
Coax
0.82 to 0.98
246 to 294
determine the value – typically, a TDR can
only identify faults below 200 Ω. The lowest
measurement possible with an insulation
tester is often around 10 kΩ, so a multimeter
may be needed to fill in the resistance
measurements between continuity (below
100 Ω) and insulation (above 10 kΩ).
If possible, use a good pair of cables running
alongside the pair under test so that a direct
comparison can be made between the good
pair and the bad pair. The fault will often
be easier to identify by looking for the
difference between the two traces. Some
TDRs offer the facility to show both traces on
the same display or to show the difference
between the two traces.
Velocity factor
When using a TDR it is necessary to tell it the
speed of the pulse in the cable. This enables
the TDR to convert the time the reflected
pulses take into distance. Different types of
cable have different velocity factors (VF).
VF is the ratio of the speed of the pulse
in the cable to the speed of light. It may
alternatively be entered as m/µs when it
is called velocity of propagation. If you do
not know the VF of the cable under test it
may be possible to test a known length and
adjust the VF until the distance displayed to
the end of the cable is correct.
Typical velocity factors
Accuracy
Open conductor
A large positive trace
Bridge tap - A small positive followed by a
small negative trace after a few mintues
A TDR cannot be used to pinpoint a fault.
The TDR accuracy will depend on the velocity
factor (maybe only with a resolution of 1%);
the cable may twist or not lie in a straight
line and the resolution of the display will not
show fine detail unless zoom is used. The
best way to locate a fault is to test the cable
from both ends; the fault will lie between
the points identified.
Output pulse level
Short circuit
A negative trace
Split/resplit - Negative trace followed by a
small negative trace after a few metres
The output pulse level can be varied to assist
in locating the fault. Small faults and those
at the far end of the cable will require a high
pulse level. High pulse energy with near-end
faults will, however, distort a large section of
the displayed trace and so lower pulse levels
will be better.
Range
Cables splice/joint
A small positive followed by a small negative
Wet splice/water
Short positive/negative trace
Initially it is best to set the range to well above
the expected length of the cable so you can
see the complete picture. Even a large fault
will be missed if it is off the display.
Automatic fault finding
T Joint - A negative trace
followed by a long positive
Water ingress
A negative trace
Many TDRs have an automatic facility that
may help identifying some faults, but it is also
necessary to be able to use the TDR manually
in order to get the most from the instrument.
ELECTRICAL TESTER - April 2014
5
ELECTRICAL
TESTER
The industry’s recognised information tool
When reflection isn’t the answer
The first is that the instrument should perform
bipolar measurements – that is, it should
automatically carry out measurements first
with one polarity and then with the reverse
polarity, and average the results.
This is necessary because, if dissimilar metals
or temperature differences are present in the
cable, offset voltages can be generated both
by thermoelectric effects and, particularly in
the presence of moisture, by electrochemical
(galvanic) effects. These effects are inherently
unipolar and the inaccuracies they would
otherwise introduce are, therefore, eliminated
by making bipolar measurements.
This technology also gives results that are
unaffected by resistance differences in the
sheath and core conductor, and by the
resistance of the auxiliary connections. In
addition, the performance of the connecting
clamps has little influence, and no special
calibration procedures are needed when
carrying out the measurements. Typical
connections for the voltage-drop method are
shown in Figure 2.
The second key factor to be considered
when choosing a high-voltage bridge is the
instrument’s discharge capability. All cables
have capacitance and, for power cables, this
typically amounts to 0.5 µF per kilometre. A
cable that is, say, 50 km long will therefore
have a capacitance of 25 µF and, when this
cable is charged to 10 kV, the energy stored will
be 1250 J. Unless the instrument can discharge
this energy safely, it is likely to sustain damage,
and there is also a significant risk of injury to
the operator. To eliminate these possibilities,
the best instruments not only have a high
discharge capability, they also measure the
cable capacitance before any measurements are
made and, if it is too high, they inhibit the test
and either reduce the voltage or even prevent
the cable being charged.
There are two points to be aware of when
selecting a high-voltage bridge for fault
location on power cables irrespective
of whether the instrument is a classical
galvanometer-based bridge or one that uses
the superior voltage drop method.
A good example of a high-voltage bridge that
makes effective use of the latest technological
developments is the new HVB10 (see Figure
3). As might be expected, this uses bipolar
measuring methods and has a high discharge
capability of 25 µF. As an additional safety
Figure 1: Cable jointing boxes with cross-bonded shields
Peter Herpertz - Product manager,
power, SebaKMT
For locating faults on power cables, pulse
reflection techniques are the most widely used
approach and, in most cases, these techniques
provide accurate and dependable results.
There are however circumstances where they
don’t work so well – for example, with highresistance faults in long paper-insulated (PILC)
cables, faults in very long cables, especially
in offshore applications, and faults in crossbonded cable systems (see Figure 1).
There are also situations where verification
is needed for the results obtained by pulse
reflection fault location techniques. This is
often the case with subsea cables where,
because of the enormous cost of deploying a
vessel to raise and repair the cable, operators
often insist on using two independent fault
location techniques to ensure that the location
has, in fact, been correctly determined.
In these situations where an alternative to pulse
reflection methods is needed, an attractive
option is the use of a high-voltage bridge.
However, conventional bridges also have their
limitations, so bridge-type instruments that
support voltage-drop fault location technology
are much to be preferred. With this technology,
the current, voltage and resistance before
and after the fault are considered in relation
to the cable length, and the test instrument
automatically carries out the necessary
calculations to display the distance to the fault,
typically in less than a minute.
Voltage-drop technology has many benefits,
not the least being that it is much less
error sensitive than bridge-based methods,
which means that the accuracy of the
results it delivers is significantly improved.
Connection principle sheath fault location
Figure 2: Typical connections for voltage drop fault location technology.
collaboration brings benefits for all
An on-going joint project between Megger
and the University at Buffalo, a flagship
institution in the State University of New York
system, is delivering major benefits for the
university and its students as well as for users
of power test equipment.
The project, which forms part of the iSEED
(Institute for Strategic Enhancement of
Educational Diversity) program and CSTEP
(Collegiate Science and Technology Program),
involves the students developing new
6
This versatile instrument offers two operating
modes. Standard mode gives good results
for typical sheath faults with resistances
up to several hundred kilohms and sheath
cross-sections from 25 mm2 to 50 mm2.
Measurements in this mode typically take
around 30 seconds.
High accuracy mode uses the full potential of
the instrument and is ideally suited for locating
difficult high resistance faults in, for example,
the inner insulation of PILC cables. This mode
also incorporates an intermittent fault detection
algorithm, and measurements – depending
on the actual cable and fault behaviour – take
approximately one minute to complete.
Designed to be easy and convenient to use,
the HVB10 has an adjustable output voltage
of up to 10 kV and incorporates an intuitive
user interface that allows all major functions to
be selected with a single turn-and-push rotary
control. Results and settings are shown on a
large high-resolution display.
For the foreseeable future at least, pulse
reflection techniques are likely to remain the
preferred method for locating the majority
of faults on power cables. Where, however,
there is a need for an alternative method,
either because pulse reflection techniques are
unsuitable or because the results produced
by these techniques must be verified, a high
voltage bridge that uses the voltage drop
method is an invaluable tool. In choosing such
a bridge, however, care should be taken to
ensure that it offers the full range of desirable
– and, in some cases, essential – features
outlined in this article.
Connection core-to-screen or core to fault location
University at Buffalo
Casey Henry
marketing program manager
feature, it not only checks the capacity of the
cable before tests are made, it also verifies the
test leads are correctly connected.
software modules for the latest state-ofthe-art relay protection test systems. While
participating in the project, the students work
under the guidance not only of the university’s
tutors, but also of Megger engineers who
have wide and current practical experience in
the relay test field and who act as mentors.
The university benefits by having access to the
latest test equipment and from the up-to-theminute input provided by the mentors, while
the students enjoy both of these benefits
plus the confidence-boosting opportunity to
produce work that will ultimately be tested
and used in a commercial environment. The
project makes additional resources available
ELECTRICAL TESTER - April 2014
Figure 3: The new HVB 10 high-voltage bridge.
thankful and appreciative to be a small part
in this experience”, said Dr Zirnheld at the
University at Buffalo.
To date, the students involved in the project
have produced software modules for testing
to develop the software modules with the
result that the end users benefit by having
faster delivery on new modules than would
be possible without the cooperation of the
university and its students.
three popular and widely used types of
“Using state of the art equipment and being
guided by industrial mentors adds value
to our student educational program. Our
students get the best of all worlds. They gain
support and knowledge from an industrial
partner; they are part of new institute on
campus; they participate in multiple programs
that immerse them into an intensive research
experience, and they have numerous
opportunities to present their work. I am
relay test software suite, and will be available
protection relay. After final evaluation and
verification have been completed, these
new modules will be added to the library of
modules that already forms part of the AVTS
for download by users of the associated relay
test equipment.
www.megger.com
ELECTRICAL
TESTER
The industry’s recognised information tool
Putting cables to the test......
Clive Pink
Product manager
Many techniques are available for assessing
the condition of underground power cables
and for diagnosing faults that occur on
these cables, but these techniques are often
presented as alternatives that compete with
each other. This is unfortunate and misleading
as, in reality, the various techniques are
complementary.
Faults on underground cables are a major
concern for every organisation involved in
the transmission and distribution of electrical
power. Such faults can have consequences
that are extremely costly and disruptive, so
it’s not surprising that there is strong demand
for test equipment that can provide accurate
information about the condition of cables and
also assist in the rapid location of faults.
A first thought might well be that this test
equipment should energise the cable at power
frequency – after all, in this way it would be
subjected to stresses that closely resemble
those it experiences when in service. There is,
however, a problem.
Cables are highly capacitive which means
that if they are to be energised continuously
at power frequency during testing, the test
set must be capable of supplying a large
amount of reactive power. A test set capable
of doing this necessarily has to be physically
large, heavy and expensive. For this reason,
power frequency testing of cables is not
commonly used.
One potential alternative is dc insulation
resistance testing, and this has many
benefits. Suitable test equipment is compact,
lightweight, moderately priced and relatively
easy to use. Typically performed at 5 kV or
10 kV, dc insulation resistance tests take just
a few minutes to carry out and, in addition
to quickly revealing major faults, they give a
valuable indication of the overall condition
of the cable. This is a very useful guide when
deciding whether the cable is fit for immediate
return to service, or whether it should be
tested further using other techniques.
A recently published article (Charles Q Su and
C R Li, IEEE Electrical Insulation Magazine,
January/February 2013) describes how, during
a five-year study, dc insulation testing was used
to decide which of a group of 6.6 kV cables
operated by a Chinese utility should be further
tested using the VLF and OWTS techniques
described later in this article.
Only 5% of the cables in the study were
selected for testing with these techniques,
but the failure rate across the whole group of
cables was nevertheless reduced by over 30%.
This clearly shows that dc insulation resistance
testing is a valid technique for determining
which cables are most at risk of failure and,
therefore, in need of further analysis.
To get the best from dc insulation resistance
testing it is important to choose the right
test set. A critical characteristic is test current
capability, as an instrument that can only
supply a small current will take a long time to
charge the cable under test, particularly if it is a
long cable, and this will unnecessarily prolong
the testing time. Market leading instruments
will typically supply 3 mA to 6 mA short circuit
current. As a rule of thumb, this will mean that
capacitive loads like cables take 2.5 seconds or
less per microfarad to charge to 5 kV.
In many medium, high and extra high voltage
substations, noise immunity and filtering is
a desirable feature. The best instruments are
capable of accommodating between 3 mA and
8 mA of noise, and filter the output in realtime to provide stable measurements. Finally,
the test set’s power source should not be
neglected. As mains power may not be readily
www.megger.com
available in locations where cable testing
must be carried out, a test set with an internal
rechargeable battery – ideally a rapid-recharge
Li-ion type – is greatly to be preferred.
While dc insulation resistance testing at
modest voltages is, as we have seen, an
invaluable and convenient first-line tool for
assessing cable condition, there are some
cases where further study is needed. These
tests most usually take the form of insulation
withstand testing at voltages higher than the
nominal working voltage of the cable under
test. In fact, in many countries withstand
testing before new cables are put into service is
obligatory, as it is part of the relevant standard.
Because of the risk of cable damage, dc testing
at these higher voltages is no longer widely
used, having been supplanted by ac very low
frequency (VLF) testing, usually performed at a
frequency of 0.1 Hz.
VLF test sets are divided into two groups –
those that apply a sine wave to the cable under
test, and those that use a cosine rectangular
(CR) waveform. Both types produce useful and
reliable results, but it is worth noting that CR
test sets are usually smaller and lighter than
similarly rated sine wave equivalents, and that
some users prefer the CR waveform as its rise
and fall times are very similar to those of a
power frequency sine wave.
VLF insulation withstand tests at 0.1 Hz usually
involve applying a test voltage of three times
the nominal working voltage to the cable for
15 minutes or, in the case of aged cables, one
hour. VLF testing therefore takes longer to
perform than dc insulation resistance testing,
but it will reliably uncover a wider range of
cable problems and will enable the majority of
“dubious” cables to be confidently classified as
either good for return to service or susceptible
to imminent failure.
Even after VLF testing, some subtle problems
may remain hidden, and detecting these is
the role of partial discharge (PD) analysis. This
involves coupling a high ac test voltage to the
cable under test and using a sensitive detector
to look for the characteristic signals produced
by PD events. Since research has shown that
PD testing at VLF using sine wave voltages does
not give good results, an alternative method
of providing an ac test voltage for the cable
is needed. This typically takes the form of
damped ac (DAC) voltage.
This works by connecting an inductor in series
with the cable under test, then charging the
cable from a high-voltage dc source. When
the cable is charged, a high-speed solid-state
switch connects the inductor in parallel with
the capacitance of the cable to form a resonant
circuit. As a result, damped oscillations at
approximately power frequency are set up in
the cable, and these provide the test voltage.
Although it is one of the more recent
additions to the family of cable test techniques,
PD analysis using DAC voltages is rapidly
growing in popularity. It is already included in
standards for cable commissioning in Spain
and the Netherlands, and is also recommended
in Germany.
Timely testing helps
restore power to 200,000
Erik Blichfeld,
Produktchef, SebaKMT A/S
When the 60 kV submarine cable between the
Danish island of Bornholm in the Baltic Sea
and Sweden was accidentally cut by a ship on
26th December 2012, the island’s 200,000
inhabitants were initially left completely
without power. However, thanks to test
data provided by a state-of-the-art cable test
van, the point at which the cable had been
damaged was quickly located, allowing divers
to effect a speedy repair.
Energi Net Denmark is responsible for the highvoltage cable network that supplies power
in Bornholm, and the company has a 24/7
contract with SE Energi to carry maintenance
and fault finding on this network.
This includes the link between the island and
the Swedish power network, which, at 48
km, is the longest high-voltage ac submarine
interconnect in Europe.
In this short article it has only been possible
to briefly consider three of the most popular
and most useful test techniques for power
cables – dc insulation resistance testing, VLF
testing and PD analysis using DAC voltages.
As we have seen each of these techniques
has its own merits and shortcomings. The key
factor to bear in mind, therefore, is that cable
test techniques are not competitive – none is
universally “better” than the others – which
means that the best and fastest results will
always be obtained by matching the test
method to the application in hand and, where
necessary, being ready to use more than one
method of testing.
Sweden
Bornholm
Island
In 2012, SE Energi purchased a Seba KMT
R30 test van and, prior to the incident,
had used the on-board Teleflex MX timedomain reflectometer (TDR) to determine the
characteristics of all submarine cables to the
Baltic islands and offshore wind farms.
As a result, when the Bornholm cable was cut,
SE Energi was able to accurately determine
the precise location of the fault by making
comparisons between new and historical
data. This was done using the travelling wave
method, with automatic distance calibration
base data stored in the memory of the TDR.
The results showed that the cable fault
was 17.4 km from the Swedish coast, and
subsequent investigation by divers revealed
that the cable, which is buried 1 m below
the seabed at this point, had been cut by a
ship dragging its anchor. With the site of the
fault located precisely and rapidly, SE Energi
was able to speedily restore normal energy
supplies to Bornholm, thereby taking the
strain off the island’s own very limited power
generation resources.
ELECTRICAL TESTER - April 2014
7
ELECTRICAL
TESTER
The industry’s recognised information tool
Q&A
Q:
For a long time, it has been
standard practice in my
organisation to carry out HV
insulation resistance tests at either
5 kV or 10 kV. Some of the latest
test sets, however, allow tests to be
carried out at 15 kV. Why is this?
A:
Test sets capable of working at 15 kV
have been produced in response to requests
from customers who, in turn, are typically
responding to the introduction of standards
like NETA MTS 1997 Table10.1, applicable to
the maximum voltage rating of equipment,
NETA ATS 2007 Section 1.5 for mediumvoltage motors, and IEC 60229 2007 for
electric cables with a special protective
function. All of these call for 15 kV testing in
One of the most widely used testing techniques in the power sector is high-voltage dc insulation resistance
testing. The principles of this type of testing are well known, but there are still aspects that give rise to
questions. Here are the answers to some of the most common of these.
some circumstances. While these standards are
relatively specialised, they do demonstrate a
trend toward higher insulation resistance test
voltages so, in addition to those who actually
need to work to these standards, there are also
some users who are specifying 15 kV test sets
as a form of future proofing.
Q: For users who don’t need to comply
with these standards, are there
any other benefits to be gained by
testing at 15 kV?
A:
As the use of 15 kV testing spreads, it is
becoming apparent that testing at this voltage
can detect many faults that are not apparent
when tests are performed at lower voltages.
Typically, these faults include fractured or
crushed insulation, damaged cable outer
sheaths, and insulation that has been degraded
by corrosive or conductive contaminants.
Q: Does insulation resistance
testing at 15 kV bring any new
safety requirements?
A:
In general, the safety requirements are
very similar to those associated with testing
at 5 kV or 10 kV, and safe working methods
should always be followed. It is also important
to ensure that the test leads used are
specifically designed to provide the extended
creepage paths associated with 15 kV testing.
And, of course, the instrument used should be
designed and constructed to offer the highest
possible level of user safety.
Q:
Why are some types of HV
insulation tester offered in versions that have different levels of noise
immunity?
A:
It’s often necessary to carry out
insulation resistance testing in areas where
electrical noise is present. If an instrument
with poor noise immunity is used in these
situations, it may deliver unreliable results or
even no results at all. To avoid this problem,
modern insulation resistance testers typically
have a noise immunity of around 3 mA – that
is, they will deliver dependable results with
up to 3 mA of noise present in the measuring
circuit. There are situations, however, where
even greater noise immunity is needed –
when testing in high voltage substations, for
instance. To accommodate these situations,
the latest testers are available in models with
up to 8 mA noise immunity. Building in this
extra noise immunity necessarily adds to the
cost of the instrument, however, so the 3 mA
models remain available as a cost-effective
alternative for users who don’t need to work
in high noise environments.
The stability factor
Damon Mount - power sales manager
Evaluating new vendors and re-evaluating
existing vendors are time-consuming yet
essential tasks. There are many factors that
need to be considered, but what are the most
important? Certainly product quality and
breadth of product range come near the top
of the list, as do value for money and quality
of customer support. But, especially when
buying products like test equipment that can
be expected to have a useful working life of
many years, there’s another factor that’s just as
important – vendor stability. Can the vendor be
relied on to provide advice and service for its
products in years to come?
One useful indicator is undoubtedly the
vendor’s past performance. Companies that
have proved themselves to be sufficiently
adaptable to cope with technology shifts
and recessionary markets in the past have
demonstrated that they have an instinct for
survival, which they’ll be able to deploy again
in the future, should the need arise. In short, if
the vendor has been around for a while, this is
a very positive sign.
However, as we’re always being told, “past
performance cannot be relied on as a guide
to future performance”. In other words, it’s
also essential to check how well the vendor is
doing right now. There are many ways to do
this – a credit agency search will, for example,
produce a lot of useful information. But, for
vendors who are among the top performers,
there can be an easier and more convenient
source of information in the form of respected
league tables, such as those published
annually by Fast Track in conjunction with the
Sunday Times.
These tables give instant access to key
information about leading companies, as
well as ranking their performance. They
can, therefore, be very useful time savers in
the corporate stability section of the vendor
evaluation process. And how does Megger
measure up? Past performance is beyond
question – the company and its antecedents
have been innovating and trading
successfully in the electrical measurement
sector for well over a century.
But what about today? A glance at the 2013
Top Track 250 league table recently featured
in the Sunday Times will confirm not only that
Megger is one of Britain’s top mid-market
private companies – an achievement it shares
with such well-known names as Dyson
and Harrods – but also that the company is
continuing to evolve and expand, it’s recent
acquisition of German cable test expert
SebaKMT being specifically mentioned.
Power problems? Bring on the mice!
Ted Kim, Regional sales manager, North Asia
and Papua New Guinea that was accidentally
introduced to the island in the late 1940s.
In most situations, if there’s a threat of power
blackouts, trying to solve the problem by
dropping mice from low-flying aircraft is not
the solution that first springs to mind. Yet
that’s exactly what the US Government is
doing in the Pacific Ocean island of Guam.
Maintaining electrical power transmission
and distribution systems in tropical climates is
always challenging, but in Guam the situation
is exacerbated by the brown tree snake, a
non-native species from northern Australia
With few natural predators and abundant
prey, the snake population exploded until
today it is estimated that there are around
three million snakes on the island. And that’s
a huge problem. They are driving many
indigenous species towards extinction and
they bite people. But that’s not all. The snakes
climb power poles to search for their favourite
“restaurants” – birds’ nests – and in doing this,
they create short circuits that lead to blackouts.
8
ELECTRICAL TESTER - April 2014
The snakes also love power stations where they
are attracted by the warmth and humming
associated with electrical power systems and
where they can be sure they’ll find rats to
provide a tasty meal. And, once the snakes
have slithered their way into power stations,
they create all sorts of havoc. The Guam Power
Authority reckons that these problems are
costing it $4 million a year.
unlikely that the snakes will ever be completely
eradicated in Guam using this technique, but
the hope is that there will a useful reduction in
the power company’s snake-related problems,
and also that the snakes can be prevented from
being accidentally transported to the island of
Hawaii, where it is anticipated that they could
cause billions of dollars of economic damage.
However, a solution may be to hand. The
snakes are almost uniquely sensitive to the
common painkiller paracetamol (also known
as acetaminophen) and just a small dose,
which would be harmless to other animals, is
enough to kill them. To take advantage of this
vulnerability, the US Government has started a
programme of dropping dead mice laced with
paracetamol from helicopters in areas where
the snakes are prevalent.
So there it is. If brown tree snakes are causing
power headaches in your locality, paracetamol
or, if you prefer, acetaminophen – might just
be the right prescription. But don’t take it
yourself – feed it to the snakes!
Animal lovers will be pleased to know that
the mice are dropped individually and, even
though it might be seen as too little too late,
each one is provided with its own parachute.
The idea is that the mice come to rest in the
tree canopy where the brown tree snakes live,
rather than reaching the ground.
Since brown tree snakes, unlike most other
snake species, are happy to eat prey that’s
already dead, the hope is that they will take
the bait and succumb to the poison. It’s
“Mice are not the only innovative resource that
Guam Power Authority uses to help maintain
the performance of its distribution systems,”
said Ted Kim, regional sales manager for
Megger. “The Authority also makes extensive
use of qualifying test products from Megger,
which range from protection relay test sets and
transformer/diagnostic equipment to revenue
metering and calibration systems.”
Credit: Parts of this story were adapted from
an original piece reported by AP which can be
found at http://bigstory.ap.org/article/us-govtair-drop-toxic-mice-guam-snakes.
www.megger.com