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