HD Electric Products

More Information

Home » Library / Archives » RF/TVI and Blinking Light Complaints: Common Symptoms of Non-Visible Insulator and Arrester Failure

RF/TVI and Blinking Light Complaints: Common Symptoms of Non-Visible Insulator and Arrester Failure

Presented at Tech Advantage 98, March 8, 1998, Nashville, Tennessee

Insulators and arresters routinely become electrically defective with no accompanying visual evidence. Historically, it has been widely assumed that defective insulators and arresters can be identified by physical evidence such as flash over burns, chipped, cracked or broken porcelain and blown disconnects. However, a growing body of data shows that just because insulators and arresters look good mechanically, does not mean they are good electrically.

Non-visible defective insulators and arresters are a common source of such expensive and time consuming operations difficulties as intermittent and persistent RF/TVI and blinking light complaints. They also pose hidden threats to line crew safety when energized maintenance work is being performed.

Data on the symptoms, patterns and costs of these non-visible defects are presented and methods of convenient and cost effective testing to identify them are described.


Periodic failure of insulators is a fact of electric utility operations. Historically, it has been widely assumed that these failures can be identified by visual inspection for physical evidence such as flash over burns and chips, cracks or breaks in the porcelain. As a result, non-visible failure of insulators is not generally considered to be a cause of operations difficulties or a threat to line crew safety.

In this section of the paper, symptoms and consequences of non-visible insulator failure at both distribution and transmission levels are described. Data on patterns and frequency of these non-visible failures are included to the extent they are available.

Nuisance Calls and Non-visible Insulator Failure on Distribution Circuits

During the last half of the 1980's and the early 1990's, Cloverland Electric Co-operative, Dafter, MI. experienced steady year over year increases in customer complaints and service reliability problems associated with:

  • radio/TV interference (RF/TVI);
  • linking lights;
  • the number of OCR operations coupled with difficulty in locating problems causing breaker operations;
  • the number of fused TAP lines requiring refusing after lightning and severe storms;
  • and pole top fires.


By mid 1991, the number of customer complaints had reached crisis levels. The amount of crew time being spent on nuisance calls was interfering with routine maintenance and planned construction activities. Furthermore, failure of the company to quickly resolve these customer complaints was having a negative impact on customer relations.

Early efforts undertaken to resolve the complaints included:

  • a concentrated tree brushing program;
  • tightening line hardware wherever RF/TVI persisted;
  • and extensive RF/TVI patrols.


None of these time consuming and costly efforts were successful in resolving the problems.

In early 1991, the company decided to investigate the possibility that non-visible insulator failure was the source of its problems. They purchased a Hi-Test Insulator Tester which would allow them to test insulators for non-visible defects while the lines remained energized (see Appendix A). Three suggestions made by the manufacturer of the tester were useful to the company in structuring its test program:

  • non-visible insulator failure has been shown to cause the problems being experienced by the company;
  • non-visible insulator failure is often more common than visible insulator failure (a cutaway example of such a failure is shown at Plate One);
  • and a large majority of non-visible insulator failures occur in suspension insulators used in dead-end applications.


From mid 1991 to mid 1994, the company tested all suspension insulators in dead-end applications on a circuit by circuit basis starting at the substations and proceeding out along the circuits. Approximately 1900 miles of lines containing over 35,000 insulators were tested

A total of 2,243 defective insulators were identified and removed from service. The type and size of these defective insulators are shown at Table One. The year of manufacture of a large random sample of the defects (approximately 75%) is shown at Table Two. In 1991, when the insulator test program was begun, the company experienced 719 outages. In 1994, the year the program was completed, the number of outages had dropped to 498 and that number was expected to drop even lower in 1995. Based on these reductions in outages, the company achieved annual savings of $75,000 on nuisance calls and a payback period on the insulator test program of slightly less than three years.

Table One


4.25" bell

6" bell

10" bell



















Table Two

Yr. of Mfg.

No. of Defects


Yr. of Mfg.

No. of Defects

Prior to 1960























































In summary, the results of this program were:


  • substantial reduction in RF/TVI complaints;
  • a substantial reduction in blinking light complaints;
  • a substantial reduction in the number of breaker operations and the elimination of a large number of time consuming and costly complaint calls;
  • a dramatic reduction in the number of fused TAP lines needing refusing after lightning and severe storms;
  • reduction of pole fire losses to virtually zero;
  • reduction of line losses from 10.42% in 1990 (the year before the program was begun) to 9.63% in 1993 (the year before the program was completed) with further reductions anticipated. NOTE: The company also operates 163 miles of transmission line and attributes at least one-third of its line losses to these circuits; and
  • savings of approximately $75,000 annually on nuisance complaints alone.


The Operations Manager at Cloverland Electric Co-op during the time of this program said, "The insulator inspection and maintenance program has had more immediate impact on service reliability than any other program I have ever adopted."

Non-Visible Insulator Failure on Transmission Circuits

The presence of non-visible insulator failure on transmission circuits is known to cause at least three operations problems:

  • unexplained ground fault trips;
  • RF/TVI complaints; and catastrophic failure of insulator strings;
  • often accompanied by the conductor falling to the ground


1) Unexplained Ground Fault Trips

Unexplained ground fault trips while not common are not unusual. Where no obvious explanation for the interruption exists and subsequent patrol of the circuit turns up nothing, it is typically attributed to birds, squirrels, loggers, firewood collectors, or weather. However, it is now known that non-visible insulator failure is a significant source of these unexplained ground fault trips. A narrative example follows:


"Until the early 1970's, this circuit had a good reliability record. At that point, during the spring and fall seasons of the next several years, the line began to experience occasional unexplained ground fault trips. The circuit would be restored by reclosing the relay and then patrolled. No apparent cause for these trips was discovered and they were generally attributed to imprudent loggers, firewood collectors, and lightning strikes. It was observed, however, that these problems were invariably encountered after precipitation. On one occasion, the circuit could not be restored and a pole fire and structure loss occurred. This event was clearly the result of insulator failure, although upon subsequent inspection there was no visible evidence of defective insulators on the structure.

In late 1975, a ground fault relay tripped the circuit and service could not be restored. Numerous attempts to reclose over a period of several hours resulted in a timed ground fault trip. After an extensive foot and helicopter patrol and just prior to darkness, service people were stationed along the line in an attempt to observe flash, noise, or any other symptom that might pinpoint the problem.

A rural customer telephoned the company to report that a pole behind his house had made a noise and he thought he had seen a flash. A service crew was dispatched to this neighborhood and another reclose was attempted. The centre string on a running vertical corner was observed to flash and was heard to be noisy.

This string was replaced and service was then restored - after a ten hour outage. Upon removal of the suspect string, a small burn mark was observed on the pole behind the curved washer. All the insulators, however, appeared normal. At this point, the company was convinced that non-visible insulator defects were causing the majority of the unexplained trips.

Every structure on the line had already been climbed and each insulator visibly inspected closely. It was concluded that a means other than visual inspection would have to be employed to 'test' suspension insulators."

Subsequent testing of the insulators on this circuit using early prototypes of the Hi-Test Insulator Tester revealed numerous 6 bell strings in which 4, 5, or all 6 bells were non-visibly defective. The subject circuit is 75 miles in length and is operated at 69 KV nominal. Replacement of the defective insulators completely resolved the problem. Prior to this insulator test program, the company was planning to replace all the insulators on the circuit - a project which would have cost in excess of $1,000,000 in their estimation.

2) RF/TVI Complaints

RF/TVI complaints associated with transmission circuits are not as common as those emanating from distribution circuits due to the difference in population density commonly found around these two types of circuit. However, they are not unknown as evidenced by the following example:


"Since 1983, more than 60 electrically punctured 9 inch diameter porcelain suspension insulators have been found on Commonwealth Edison's (CECO) 345 KV transmission system as a result of television interference (TVI) complaints from customers living near R/Ws ... (we) suspect the problem is more widespread than otherwise indicated by the number of TVI customer complaints. Many miles of transmission lines are located in rural areas far away from residential developments. Therefore, we would not expect to receive TVI complaints in these areas if punctured insulators were present."

RF/TVI complaints associated with non-visible failure of suspension insulators used in strings will be intermittent in nature. The critical element is moisture. As long as the external surface of the insulators in the string remains dry, the punctured insulator(s) will not 'arc'. However, once the external surfaces of the insulators are damp enough to become conductive, those insulators which are internally punctured will 'arc' thereby causing the RF/TVI. It is this intermittent, weather dependent nature of the problem which makes location and identification of these defects so problematic.

3) Catastrophic Failure of Insulator Strings

Catastrophic failure of a string of insulators characterized by a high proportion of the insulators being badly damaged or totally destroyed is described in the following example:

"In October, 1984, a live line crew was called out to change a dead-end string of insulators which were initially believed to have been damaged by rifle fire. Only five unbroken bells remained on the string, the other seven having shed their porcelain. Close examination of the fragments on the ground indicated they had not been shot as the porcelain skirts were broken almost exactly in half and, in some cases, much of the bonding material was still attached to the porcelain fragments.

In this company, suspension insulators are routinely tested to determine that they can be safely hot sticked before any live line procedures take place. Following normal operating procedures, the crew tested the string using the Hi-Test Insulator Tester and discovered that only three insulators (one of which was a cob) were electrically effective. The other nine were demonstrated to be defective by the tester. Testing of the neighboring insulators on the structure (all of which appeared good) located a further five defective insulators."

The common characteristics of this type of failure include: progressive physical damage along a string of insulators ending with an insulator for which only a metal cap remains; charred, blackened and/or twisted metal caps and pins on one or more insulators; separation of the insulator string resulting in the conductor being dropped; and pieces of the porcelain with bonding material attached to them lying on the ground beneath the subject string.

The mechanism for such failures is hypothesized to be non-visible defective insulators in the string in conjunction with moisture inside one or more of these defects. If a string in this condition flashes (as a result of lightning or a switching surge), then the moisture inside the insulator(s) would be vaporized and the insulator(s) blown apart. The explosive force originating inside the insulator would explain the pieces of porcelain on the ground with bonding material attached to them.

Patterns of Non-visible Insulator Failure

Non-visible insulator failure is not distributed randomly across the system. Data collected from the field over the past ten years shows two strongly non-random patterns to these failures which pose hidden threats to line crews performing energized work.

  1. They occur primarily on dead-end structures (see Table Three)

    There seem to be three reasons for the concentration of these defects on dead-ends: mechanical, electrical and weather stress. Insulators installed on dead-ends are known to be subject to more mechanical and electrical stress than those installed on tangents or running corners. In addition, insulators installed on dead-ends have the bonding material much more exposed to wetting from rain or snow than those installed on tangents or running corners. This moisture is known to invade the bonding material of some insulators and can cause internal pressure in those insulators in two ways: i) by producing a chemical change in the bonding material, known as cement growth; and ii) moisture present in the bonding material will expand when subjected to freezing weather; this pressure can be relieved through a crack in the porcelain head of the insulator.
  1. Table Three



    No. of Insulators Tested

    No. Of Defects Found

    Prob. of Defect

    Running corner




















  2. They "cascade" within a string of insulators (see Table Four)

    The reason(s) for the "cascading" pattern of failure is not known. However, the implications for line crew safety are clear from the data in Table Four and the discussion of that data which follows.

    Table Four


    No. of Non-visible Defects

    No. of Strings

    Prob. of Defect

    1 or more



    2 or more



    3 or more



    4 or more



    5 or more



    6 or more



This data is from a test program in which a total of 414 strings of insulators (with 12 bells per string) were tested for non-visible insulator failure. Of that total, 115 strings (or slightly over one-quarter of them) had at least one non-visible failure. Furthermore, once a string has been found to have a non-visible defect, the likelihood of finding additional defects in that string was more than double the likelihood of finding the first defect in the next string to be tested.

Insulators were tested on a total of eighty-four structures in this program. Examination of the data structure by structure identified forty-two (i.e. one-half) which required maintenance attention to one or more insulator strings. Examination of the data from dead-end structures only, identified forty-one of the fifty-four structures tested (i.e. three-quarters) requiring maintenance attention to one or more insulator strings looked at another way, 41 of the 42 structures needing maintenance attention to insulators where dead-ends. Finally, of the 115 strings in which non-visible failures were found, 31 had levels of insulation which were reduced by one-third or more.

At the time of this test program the circuit was operating with no interruptions or other symptoms of these defects. These data graphically demonstrate the hidden threat to line crew safety which can be produced by these two patterns of non-visible insulator failure.


As with insulators, operations and maintenance personnel often assume that an arrester which looks good is good. This is generally coupled with the belief that a defective arrester will identify itself quickly either through the operation of the disconnect or by way of a fragmentation failure (in the case of porcelain housed arresters).

However, arresters do not always identify themselves quickly once they become electrically defective and there is evidence that non-visible failure of distribution class surge arresters are also a source of blinking lights and RF/TVI complaints.

These difficulties are frequently treated as transient problems and the service procedure is often to wait for the problem arrester to identify itself through operation of the disconnect or by fragmentation failure. In areas of high lightning activity, where the number of arresters installed across the system is large, such a maintenance strategy can produce circumstances where operations difficulties associated with non-visible failure of arresters become an on-going problem.

In this section of the paper, types of arrester failure are discussed with particular attention given to those types of failure which occur gradually. It is these gradual failures which produce the operations difficulties described above. Where available, data on the causes of these failures are presented.

Pathways to Distribution Class Arrester Failure

Three types of distribution class arresters are commonly employed today:; i) the metal oxide (MOV) type which consists of metal oxide blocks in series between line and ground with no spark gap spaces; ii) a recent variant on the MOV type which includes resistance gaps together with metal oxide blocks; and iii) the older silicon carbide type which consists of a series of silicon carbide blocks and spark gap spaces. Both metal oxide and silicon carbide blocks have non-linear resistance such that they are highly resistant to electric current up to a specified voltage and then very rapidly become conductive as that voltage is exceeded.

Failure of any of these types of arresters can occur either suddenly or gradually.

A)  Sudden failure occurs for the following reasons:

  • puncture of the blocks due to heat channeling caused by excessive energy release from either a single large surge or repeated operations in rapid succession;
  • internal flashover of the blocks due to contact arcing;
  • internal flashover of the gaps and valve elements due to moisture invasion; or
  • dielectric failure of the collar material of the blocks.


Sudden failure of distribution arresters does not pose nearly the degree of operations difficulty or maintenance expense which gradual failures pose. Sudden failures are generally accompanied by ample physical evidence of the failure either through operation of the disconnect or fragmentation of the arrester housing. As a result, the time and cost of replacement is generally quite small since they are quickly and easily located (assuming there is no personal injury or property damage if the failure is a fragmentation type).

 B) Gradual failure occurs for the following reasons:

  • flashover of the spark gap spaces due to moisture invasion or metallization of electrode spacers due to heavy surge currents;
  • repeated sparkover of gap spaces due to external contamination or repetitive switching surges;
  • failure to interrupt the power follow curve;
  • operation in the presence of abnormal voltage overload;
  • thermal runaway of gap assembly grading resistors (silicon carbide type) or excessive energy absorption, block element degradation or abnormal system voltages (metal oxide types);
  • dielectric failure of one or more of the block elements (metal oxide types); or
  • irregularities at the surface interface between blocks; or
  • minor moisture invasion creating a conductive path parallel to the block elements (metal oxide types).


Gradual failure of arresters can take anywhere from several days to several weeks to reach a point where blown disconnects or fragmentation failure occur. In high lightning activity areas, where the number of arresters installed is large, gradual failure of arresters can produce ongoing blinking lights and RF/TVI complaints. The absence of visual evidence of the defect in the field makes traditional trouble shooting activities cost in-effective. The Hi-Test Surge Arrester Tester allows maintenance personnel to address these problems in a cost-effective manner (see Appendix B).

Causes of Arrester Failure in the Field

Collection of failed distribution class arresters over several months by field personnel at Ontario Hydro produced several hundred arresters subsequently examined by R&D staff at Ontario Hydro. Their study showed that these failures were due to the following causes:

  • moisture invasion - 86% of failures
  • lightning - 6% of failures
  • surface contamination- 5% of failures
  • misapplication - 2% of failures
  • unknown - 1% of failures


The arresters included in this study were all porcelain housed, silicon carbide types. The results show very clearly the sensitivity of these arresters to moisture invasion. It requires the smallest failure of an end seal gasket to allow air to pass in and out of the arrester as ambient air and arrester temperatures change. Extremely small quantities of moisture inside arresters produce failure. Such moisture invasion failures can occur even on clear days as the failure is ultimately triggered by the redistribution of the moisture inside the arrester due to temperature changes in the air. Frequently such failures are triggered by the thawing of frozen moisture in the arrester or re-energization of the arrester during maintenance work.

The shift to polymer housed arresters and metal oxide technology is a method for reducing the air space inside arresters thereby reducing the likelihood of moisture invasion. Widespread field application of polymer housed MOV arresters will change the pattern of failure away from moisture invasion toward thermal problems as the most common cause of failure.

This will not change the operations difficulties experienced in the presence of gradual, non-visible failure of arresters. Blinking lights and RF/TVI are both complaints known to be caused by gradual failure of arresters for thermal reasons.


Non-visible failure of insulators and surge arresters are sources of operations problems which are difficult to locate in the field and costly to the utility. RF/TVI and blinking light complaints are common symptoms of the presence of these non-visible failures on distribution circuits.

Data showing the frequency of these failures and some of the non-random patterns which are characteristic of these failures are presented and discussed. Attempts to resolve these problems through traditional trouble shooting activities are generally non-productive and costly.

Two pieces of equipment developed specifically to assist in the identification of non-visibly defective insulators and arresters are described.




This piece of equipment was developed in response to the types of service reliability problems and threats to line crew safety posed by non-visible insulator failure described in this paper. It allows line crews to test insulators quickly, safely and reliably while they remain in service and the line remains energized.

It includes the following features.

  • the tester can be used on any AC circuit from the lowest distribution to the highest transmission voltage with the circuit either energized or de-energized;
  • it fits on the end of a standard hotstick and is light enough for any lineman to use unassisted - weight is approximately 2.7lbs or just over 1kg;
  • it is a self contained DC power source which is unaffected by high AC voltages imposed across the test probes, thus allowing a lineman to "meg" insulators in service;
  • it imposes a 10 KVDC potential on any object placed between two probes extending from the rear of the tester and indicates the condition of the insulator by way of an LED display and an audible warning buzzer on the front of the tester;
  • the 10 KVDC potential is current limited so that no shock hazard is present for the user and the test is non-destructive to insulators; and
  • it is powered by a rechargeable 8 VDC battery with the recharger built into the tester.




This piece of equipment was developed in response to the problem of gradual failure of distribution class surge arresters described in this paper. It can be used either in a stores yard/warehouse setting to test materials being recycled or it can be taken to the field and used in a trouble shooting activity to locate non-visible defective arresters which are causing operations difficulties.

It includes the following features:

  • it imposes a current limited voltage on the arrester under test;
  • the voltage imposed on the arrester can be varied continuously from zero to the maximum output of the tester;
  • it displays the voltage withstand capability of the arrester under test in both actual voltage DC and the RMS equivalent AC voltage;
  • it displays the current leakage of the arrester under test in microamps;
  • the test is non-destructive to arresters;
  • it is equipped with both a pivoting handle and a belt clip for convenient portability and ease of use in the field (weight is approximately 1 kilogram or a little over two pounds); and
  • it is powered by a 12 VDC rechargeable battery with the recharger built into the tester.


John A. Farquhar, Ph.D
President, Hi-Test Detection Instruments Inc., Blaine, WA.
Copyright, 1998


Click here to visit the Insulator Tester product page

Click here to visit the Surge Arrester / Leakage Tester product page

Click here to return to Library/Archives page

Back | Top