Reducing Ground Noise Emissions in Critical Electronic Equipment and Facility Ground Systems to Help Prevent Conducted EMI Problems

Many EMI problems involve emissions that circulate through electrical conductors within a facility and within equipment

Philip F. Keebler and Kermit O. Phipps
EPRI
Knoxville, Tennessee

Background
The electromagnetic environment includes the areas inside and outside electrical and electronic equipment. The areas inside equipment include unused air spaces, circuit boards, cables and wires, and passive and active semiconductor components. The areas outside equipment include mounting and suspension systems for equipment, the bricks and mortar of all types of facilities, the heating-ventilation-and-air-conditioning and plumbing equipment, the power conductors that tie the equipment together and the equipment to the facility power system, other equipment inside a facility, and the earth beneath the facility floor and the air above the roof. Within this mix of areas and spaces, one part of any electromagnetic environment (arguably one of the most critical parts), lies a complex system of “zero-potential” conductors that literally (or attempt to) tie all electrical systems and equipment together. This system is called the ground system.

Ground systems play a critical role in electrical safety, power quality, system compatibility, equipment and circuit protection—and, just as importantly, the management of conducted emissions. Ground systems provide a path for fault currents to flow regardless of how the fault current was generated. Without ground systems, circuit protection devices (e.g., fuses and circuit breakers) could not function properly; and humans would be at an increased risk of electrocution. Fault currents would cause more equipment damage. The surge currents generated by the operation of surge protective devices (e.g., metal oxide varistors) would have no place to flow.

As more applications of high-frequency electronic end-use loads are developed, the amount of electrical noise injected into the grounds of facility electrical systems is increasing drastically. In many facilities, ground noise currents have reached levels that are threatening to the uptime and reliability of electronic control equipment used in industrial, manufacturing, and commercial processes. End-use equipment designers attempt to combat power line noise by including conducted line emissions filters on the line inputs of equipment. Line emissions filters are designed to reduce line noise to levels as required by product standards promulgated by the Federal Communications Commission (FCC), International Electrotechnical Commission (IEC), CISPR, and CENELEC among other standards bodies. Although many of these filtering efforts are successful, part of the filtered noise is directed into the ground circuit by design, resulting in an increase of noise on the ground circuits of facilities.

End-use equipment that depends upon the same ground circuit may be a sink for some of the ground noise currents at frequencies where the impedance may be low enough to allow selected noise currents to enter critical equipment. Noise currents that enter critical equipment may also cause equipment to malfunction and/or shutdown.

Ground noise filters may be used to reduce ground noise currents. Third-party research and testing with respect to the electromagnetic compatibility (EMC) performance of these filters under various voltage and impedance conditions, combined with their application in different end-use equipment scenarios, was carried out to illustrate:

• The variation of filter insertion loss with impedance and load.
• The variation of filter attenuation performance with impedance and load.
• The reliability of the filter technology under normal (and abnormal) electrical environmental conditions.
• The limitations of the filter technology with respect to voltage quality, ground noise levels (and bias conditions), and potentially harmful currents such as surge currents, fault currents, etc.

Introduction
Circuit-based solutions to power quality (PQ) and electromagnetic compatibility (EMC) must be able to survive the effects caused by electrical disturbances and even adverse noise conditions. The degree of filtering effectiveness is linked to the presence of electrical disturbances and transient currents that must pass through ground circuits to maintain safe operating conditions while still providing some level of acceptable ground noise filtering. Through the application of compatibility engineering evaluations, the EMC performance—insertion loss and attenuation performance—of filters such as ground noise filters can be determined as a function of normal and adverse electrical conditions—real conditions that exist in customer facilities of all types. Without compatibility engineering evaluations, filter manufacturers assume the risk of providing their customers with filtering solutions with performance that cannot be verified under the real world electrical conditions characteristic of commercial and industrial facilities. Such evaluations help both manufacturers and end-use customers to gauge the robustness of filtering solutions.

Approval of electrical and electronic devices, including protection and filtering devices, by Underwriters Laboratories (UL) is based on safety, not power quality or electromagnetic compatibility. In filter applications, whether they are intended for power-line or ground filtering, reliability depends upon the ability to maintain insertion loss and attenuation performance when the filter is placed in the normal electrical environment. Significant changes in insertion loss and attenuation performance should occur only as a result of normal component (resistor, capacitor, inductor, etc.) and assembly age. Such changes should not come about as a result of real world impedance variations (source or load) including, and combined with, electrical disturbances, and filter loading effects. Several filter manufacturers specify no load insertion loss or attenuation performance, seemingly not realizing how load currents across the frequency band of operation affect filter performance—even as filter noise loading increases. In most specifications and applications of this type, insertion loss and attenuation performance collapse as a result of linear and non-linear loading.

Identifying the robustness of filter solutions is critical to predicting filter performance in real electrical environments—environments that are not perfect and never will be. Ground circuits were designed to provide a path for transient currents whether they result from surges, faults, or other “normal” conditions. With the application of various types of surge protection devices (SPDs) in end-use equipment and the introduction of power electronic devices of higher frequency, the opportunity for these currents to flow through the ground circuits is increasing—a trend that highlights the importance of carefully defining the performance of filtering solutions under transient current conditions. This testing program has been designed to address these issues for ground filter products.

Utility Customers Experience Ground Noise in Critical Applications
One of the functions of EPRI’s Power Quality Program is to maintain and provide a Power Quality Hotline that is accessible to EPRI-member utilities for advice on power quality and electromagnetic compatibility problems associated with utility operations. Utilities can then provide advice on facility and end-use equipment operations to their customers. Facilities with power quality and electromagnetic compatibility problems include residences, commercial facilities, manufacturing plants, industrial plants, Internet and corporate data centers, and health care facilities, among others.

Each year EPRI’s Power Quality Hotline receives several hundred calls regarding power quality problems and equipment malfunctions. The availability and general performance of power-conditioning and filtering solutions familiar to EPRI’s power quality and electromagnetic compatibility engineers is typically shared with utilities and is then provided by the utilities to their customers. A utility power quality investigator might ask, “How would one go about solving a conducted electromagnetic interference (EMI) problem with a piece of end-use equipment?” Additional information can be provided to utilities and end users regarding the solutions (i.e., the general types of equipment that could be used to solve a problem) that have been tested for performance by EPRI. The range of problems called into the Hotline varies from equipment shutdown to equipment malfunctions. Ground noise problems are among those reported problems. In fact, non-destructive equipment shutdowns and malfunctions that are not directly attributable to ground noise might, in fact, be related to ground noise problems if all of the facts were known.

In many cases, equipment problems reported through the Hotline are further addressed by EPRI through power quality and electromagnetic compatibility facility audits and investigations. In these efforts, compatibility engineers comb through the wiring and grounding systems that support end-use equipment operations looking for wiring and grounding (and other) problems that contribute to equipment shutdowns and malfunctions. The measurement (time and frequency domain) of ground noise currents at the facility, electrical panel, and equipment level are among the many measurements included in these efforts. As a part of such efforts, compatibility engineers may discuss in general the specific types of solutions, including those that have been tested with verifiable and documented performance. These solutions can then be considered by utilities and their customers to resolve various problems including equipment upset and malfunction caused by and related to ground noise problems.

Traditional Approaches to Solving Conducted EMI Problems with End-Use Equipment
Electromagnetic interference (EMI) problems can be caused by radiated emissions, by conducted emissions, or by a mix of radiated and conducted emissions. Radiated and conducted emission mechanisms work together in magnitude and frequency and in coupling. Although filters and shields are commonly used as solutions to EMI problems, these applications tend to rely on reducing the electromagnetic energy at the source of the emissions by altering the path between the source and the receptor, or by increasing the immunity of the equipment affected—not by involving any part of the grounding system.

When addressing conducted EMI problems, efforts tend to be placed on the lowering or eliminating the emissions on power and signal conductors. Emissions traveling on any metallic media, including power and signal conductors “hunt” and seek out the area of a circuit with the lowest impedance at a certain specific frequency and/or group of frequencies. Conducted emissions actually circulate through a ground system depending upon the emissions frequency and the state of the impedances on the ground system.

EMI investigators rarely consider the characteristics of the grounding system when conducting an EMI investigation, or when trying to identify a solution to an existing EMI problem. Grounding conductors are thought of as providing low-impedance paths for fault and surge currents, and that is exactly what they are supposed to do. Still, ground conductors can also provide a “backdoor” for emissions to enter equipment and an “unguarded exit” for conducted emissions to leave equipment.

Conducted emissions that are present within a grounding system circulate throughout the system. This circulation path includes end-use equipment, especially the types of equipment that have more than one ground. Also, it is not uncommon for circulating emissions on a ground system to involve power, neutral, and signal (or data) conductors. The passive component that ties the ground system to the power, neutral and other conductors (i.e., signal and data) is the EMI filter capacitor that has one leg tied to ground. (In some instances, other passive components such as high-Ohm resistors are used in this circuit.)

Emissions that circulate through a ground system constantly change magnitude, shape, and frequency. Moreover, with the increasing number of new emissions-generating end-use devices being placed in facility electrical systems each year, the state of emissions in the grounding system are constantly increasing. In most cases, emissions in a grounding system do not cause an EMI problem; however, when such a problem occurs, it is usually a serious one that may or may not be resolved by traditional solutions or approaches. Also, these increasing levels of ground emissions present increasing levels of risk to electronic equipment on the facility electrical system and an increased likelihood that an EMI problem will occur.

Another aspect of increasing emissions on ground systems is the number of end-use devices in a facility that a) are being affected in some way by emissions entering the device’s ground and b) are on the verge of experiencing a serious EMI problem. It is a well-known fact among EMI investigators that many malfunctions and shutdowns of end-use equipment involve the magnitude and frequency of emissions traveling on the ground system.

Laboratory Testing
The compatibility engineering evaluation tests described below were carried out using recently developed and improved filter measurement methods (published in the new IEEE Standard 1560, “Standard for Methods of Measurement of Radio Frequency Power Line Interference Filter in the Range of 100 Hz to 10 GHz”) developed by the IEEE. Although these methods were developed to test power-line filters, their basic concepts can be applied to ground noise filters to determine their true attenuation and insertion loss performance. These new methods were developed for the accurate documentation of filter performance in filtering devices that operate in a wide range of frequencies and at specific frequencies. To assure effective testing, all of the tests listed below were carried out with dedicated filter samples taken from off-the-shelf stock.

Four models of ground noise filters were tested—a 20-Amp filter, two 5-Amp filters, and a one-Amp filter. The 20-Amp filter was designed to resolve ground-related EMI problems associated with certain size electrical panels. One of the 5-Amp filters was designed to resolve EMI problems with small electrical panels and end-use equipment such as adjustable speed drives, building control systems, and other noise-producing electronic loads. The second model 5-Amp filter was designed to resolve EMI problems in the broadcasting and recording industry. The 1-Amp filter was designed to resolve EMI problems associated with low-current ground problems including those found inside end-use devices.

Test Protocol
Small-signal-type bench top tests are necessary to determine the basic EMC performance of the ground noise filters under laboratory controlled conditions. Insertion loss and attenuation performance tests are critical in the identification of the performance of end-use electronic loads including both protective and filtering devices, especially when the objective of such devices is to reduce electrical/electronic noise (i.e., conducted emissions). Insertion loss and attenuation performance tests can be carried out on many type of filters to determine baseline performance and then can be repeated after altering circuit characteristics (e.g., circuit impedances, levels of incoming noise, circuit loading, etc.). These basic tests were selected from the new IEEE 1560 Standardand then were modified to test ground noise filters. Then, in repeat testing, electrical stress is placed on end-user equipment by applying common-mode electrical disturbances (or events) to the filters. This process determines which electrical events alter the electrical and electromagnetic performance of the filters.

Two groups of tests were performed on the filters: non-destructive tests and destructive tests. The non-destructive tests were carried out to determine the basic attenuation and insertion loss performance of the filters. During these tests, attenuation and insertion loss performance can be identified for no-load ground circuit conditions under balanced and unbalanced impedance conditions and for loaded ground circuit conditions under balanced and unbalanced impedance conditions for various frequency ranges.

Power quality and EMC investigators may find that ground circuits carry various levels of ground current at 60 Hz and at other frequencies. Without a doubt, they will find that the lengths of power and ground circuits vary along with the total power and ground circuit impedance. (Ground noise currents also flow in a closed loop; thus the power and ground impedances must be considered.) Ground impedances may be more than an Ohm or two, or more than that allowed by the National Electrical Code. Increases in ground impedance may be caused by poor bonding, ground conductors that are sized too small, and other wiring and grounding errors and conditions. Variations in the magnitude and frequency of ground currents and ground circuit impedance will impact the performance of ground noise filters. Consequently, Tests 1 through 5 must be carried out.

Within the non-destructive test set, five types of insertion loss and attenuation performance tests were selected to determine how the insertion loss, attenuation performance, and reliability vary with the application of these electrical events to the ground noise filters. The six tests required in the compatibility engineering studied were as follows:

• Test 1 – Quality Assurance Testing—no load: (10 kHz to 1 GHz)
• Test 2 – RF Characteristic Test, mismatched ground impedance—no load (100 kHz to 30 MHz)
• Test 3 – Variable Ground Circuit and EMI Filter Impedance Attenuation Measurement—loaded (100 Hz to 100 kHz)
• Test 4 – Attenuation Measurement—loaded (100 kHz to 30 MHz)
• Test 5 – Inductor Saturation Verification Testing—loaded: (100 Hz to 100 kHz)
• Test 6 – Audio Hum Test (20 Hz to 20 kHz)

Because ground noise filters are placed within the ground circuits, investigators need to know if and how transient ground currents will affect the filters. Ground circuits are required to provide a low-impedance path for high currents (or fault currents) resulting from faults occurring at various system levels. If a low-impedance path is altered, the flow of fault currents may also be altered. Attenuation and insertion loss tests were performed before (Test 1 was used as a baseline) and after the respective destructive test was carried out. Because the application of ground noise filters has been shown to resolve audio interference and hum problems in the broadcasting and recording industries, Test 6 was conducted as a general performance test.

Test 7 was designed to determine if typical fault currents could present conditions of increased impedance and could, therefore, alter filter performance. Because ground circuits must also support currents resulting from voltage surges, Tests 8 through 10 were designed to determine if surge currents could present conditions of increased impedance and could, therefore, alter filter performance. Because the application of a ground noise filter (depending upon its location in the ground circuit) may reduce electrostatic discharge currents, Test 11 was conducted as a general performance test.

• Test 7 – Fault Current Test
• Test 8 – 1.2/50 µsec–8/20 µsec: Combination-Wave Surge Test – Individual and Repetitive Surges
• Test 9 – 0.5-µsec, 100-kHz Ring Wave Surge Test – Individual and Repetitive Surges
• Test 10 – 10/1000-µsec High-Energy Deposition Wave Surge Test
• Test 11 – Electrostatic Discharge (ESD) Protection Effectiveness Test

For brevity, only the results of Tests 1, 2, 7, and 8 will be discussed in this article.

A Brief Discussion of Test Results
Test 1: What is a Quality Assurance Test?
Quality assurance testing can be used to gauge a range of performance criteria. In testing electrical filters, it can be used to determine if the filter is functional before other tests are applied. It is also used to generate basic baseline performance data for comparing the results of other tests that may damage and/or alter the performance of filters. For example, if a filter is subjected to 1000 applications of a severe voltage surge, then the filter’s ability to attenuate noise can be determined after comparing the attenuation performance of the surged filter to the attenuation performance of a virgin filter. The frequency range of 10 kHz to 1 GHz is selected because this includes the frequency ranges of the other tests and extends to 1 GHz for identification of filter performance issues including natural and induced (resulting from damage to a filter) resonance points.

Compatibility engineering evaluation under Test 1 was used to determine the performance of the ground noise filter under no-load conditions from 10 kHz to 1 GHz. This test was conducted to establish a “baseline” performance of each filter tested before that specific filter was subjected to the remaining tests. To meet the critical requirements of this test, well-matched 50-Ohm laboratory-grade impedances were used. This test was used to determine the condition of the filter to be tested to avoid testing a filter that might already have component problems or a defective assembly.

This test was based on the no-load 50-Ohm matched impedance method, defined in IEEE 1560, and was used for quality assurance filter testing to verify the insertion loss characteristics of sample filters from 10 kHz to 1 GHz under standard 50-Ohm impedance conditions. The purpose of this test was to uncover possible component malfunctions or defective assembly. This testing is extendable in frequency range, if preliminary results indicate that the frequency range should be extended in either the low or high-frequency direction.

Results of Test 1
Each of the four filters was subjected to Test 1 to determine their baseline filter performance. Three filters of each model were included in the test. This test was carried out according to the specifications described for Quality Assurance Testing in IEEE 1560. Figure 1 illustrates the results of this test for all four models of filters. One can see that the attenuation performance is approximately 30 dB from about 100 kHz to a few MHz.

Test 2 – RF characteristic test, mismatched ground impedance—no load (100 kHz to 30 MHz)
The ground impedances found in some facilities are also mismatched or become mismatched with the age of the grounding system. It is critical that real-world ground impedances (that will be mismatched, whether looking into the end-use equipment or into a facility ground) be included in this test. Specific real-world mismatched bench-top impedance characteristics in the range of 0.1 to 100 Ohms representative of selected electronic systems (adjustable speed drives, programmable logic controllers, power supplies, and semiconductor fabrication equipment) are used in the test circuit. Compatibility engineering evaluation under Test 2 will determine the insertion loss, a critical noise filter performance metric, of the selected ground noise filter under no load conditions from 100 kHz to 30 MHz. This test. based on the no-load mismatched impedance method defined in IEEE 1560, should be used for model verification measurement to determine the insertion loss characteristics from 100 kHz to 30 MHz under specified mismatched impedance conditions.

Figure 2 is provided as an example of one of the tests in IEEE 1560 that was modified to determine the effects of mismatched ground circuit impedance conditions. The ground filter under test is the “filter” in Figure 2. All of the equipment used in the test setup is basic EMC measurement equipment with the exception of the balluns shown on the input and output of the filter. The 50-Ohm to 0.1-Ohm ballun and the 100-Ohm to 50-Ohm ballun represent the mismatched ground impedance conditions. (Because the balluns required for this test were not readily available as standard EMC test equipment, the model NH16434 and NH16435 balluns were provided by North Hills Signal Processing as a contribution to this research.)

This test is based on the current-injection method, defined in IEEE 1560, over the frequency range of 100 Hz to 100 kHz used to measure the attenuation of the powered filter under a realistic, non-50-Ohm impedance condition. The current-injection method is used to demonstrate the actual attenuation of the filter under different load conditions for linear and non-linear loading. The current was injected on the load side in this test because these ground noise filters are symmetrical filters. With this method, an additional series line reactance is used in the test over the frequency range for worst-case design testing and verification.

Results of Test 2
As an example, the results of this test for the basic 5-Amp filter will be illustrated and discussed.

Mismatched Condition:
Source Impedance = 0.1 Ohms and Load Impedance = 100 Ohms. Figure 3 illustrates the attenuation performance for the 5-Amp Ground-Nite filter under a 50-Ohm to 50-Ohm matched impedance condition versus a 0.1-Ohm source impedance and a 100-Ohm load impedance. As expected, one can see that the 50-Ohm matched impedance condition presents a better attenuation performance near 30 dB. The mismatched impedance condition case reduces the attenuation by approximately 15 dB across the entire frequency response with the exception of the frequency band near 10 to 50 kHz where the difference is closer to 10 dB. Although this is an expected result, the degree of mismatched impedance in a power and ground circuit is not likely to be this severe. Despite the reduced attenuation caused by mismatched impedance conditions, the overall result indicates the efficacy of the filter in combating ground-related EMI problems.

Mismatched Condition: Source Impedance = 100 Ohms and Load Impedance = 0.1 Ohms. Figure 4 illustrates the attenuation performance for the 5-Amp Ground-Nite filter under a 50 Ohm-50 Ohm matched impedance condition versus a 100-Ohm source impedance and a 0.1-Ohm load impedance. As expected, one can see that the 50-Ohm matched impedance condition presents a better attenuation performance near 30 dB. The mismatched impedance condition case reduces the attenuation by approximately 10 dB across most of the frequency response. At frequencies below 175 kHz, the filter has a better attenuation performance under mismatched conditions.

Test 7 – Fault Current Test
Fault currents flow as a result of a fault or unstable electrical condition. Grounding systems are put into place to provide a path for fault currents to flow. If there is no ground providing a path for the flow of the fault current—at least for the period of time needed for a circuit breaker to operate, then an unsafe condition may result during or after the fault. This test was designed to uncover the existence of a number of unsafe conditions. Specifically, a fault current of 5,000 Amperes is injected into the ground of an electronic system with a ground noise filter installed in the ground path. Testers then checked for any damage to the filter components or filter assembly, as well as for the following unsafe conditions:

• Emission of flame, molten metal, or glowing or flaming particles through any openings (preexisting or created as a result of the test) in the filter.
• Charring, glowing, or flaming of the supporting surface.
• Ignition of the enclosure.
• Creation of any openings in the enclosure that result in accessibility of live parts.
• There shall not be evidence of degradation or separation of a trace from the printed-wiring board of the filter.

Fault currents beginning at 1000 Amperes in steps of 1,000 Amperes are injected into the filters until 5000 Amperes is reached. This approach helps to identify the maximum fault current that the filters can tolerate and the degree to which their attenuation and insertion loss are affected by a fault current.

Fault current tests were carried out by another testing organization to determine if the filters met the UL 1283 standard. However, those tests were carried out to determine the safety performance of the filters and did not reveal the characteristics of the insertion loss performance after the fault current occurred.

The following requirements must be met during this test. Off-the-shelf samples of the ground noise filters must be obtained for use in this test. Fault current were injected only into the grounding circuit of the electronic system. The circuit included the expected length of the grounding conductor through the filter (elements) and the return into the remaining length of the grounding conductor. A maximum fault current closest to 5,000 Amperes was utilized as the fault current value, but higher fault currents were generated and used in this test. Filters that were fault current tested were subjected to Test 1 again to determine if their baseline performance had changed, and, if so, by how much (dB) versus frequency. Filters that experienced physical damage were documented and some of these results are included here as well.

Results of Test 7
Figure 5 illustrates the insertion loss performance of the 5-Amp Ground-Nite filter before and after a 1500-Amp (Figure 5) and a 4500-Amp (Figure 6) fault current was injected into the filter. One can see that the insertion loss for the filter exposed to the fault current was slightly affected, but the general insertion loss performance was not altered.

Test 8 – 1.2/50-µsec-8/20-µsec: Combination-Wave Surge Test – Individual and Repetitive Surges
High energy surges such as those caused by lightning and high-voltage switching are known to exist in low-voltage power systems and can cause failure of electrical and electronic loads. The 1.2/50-µsec, open-circuit voltage part of the Combination Wave, described in the ANSI/IEEE Standard C62.41.2-2002, “IEEE Recommended Practice on Characterization of Surges in Low-Voltage (1000 V and Less) AC Power Circuits,” has long been used to represent low-voltage surges caused by lightning on overhead lines. A corresponding 8/20 µsec short-circuit current waveform has also been defined with levels appropriate to the typical locations of end-use loads in facilities and can be extended to the use of filters on power line and in ground circuits. These two waveforms have substantial energy-deposition capability, provide representative stresses to the device-under-test, and can cause failures of electronic loads and associated protection and filtering devices connected in either line, neutral, and/or ground circuits. The peak values of the applied Combination Waves for ground filters are a 1.2/50-µsec, 6-kV open-circuit voltage surge and an 8/20-µsec, 3-kA short-circuit current surge.

This test is designed to determine the susceptibility of ground noise filters to Combination-Wave surges administered to the filter, first, in a non-repetitive fashion. It is desirable for a noise filter to operate without upset from these surge events. Non-repetitive surge testing beginning at 500 Volts and reaching 6 kV in steps of 500 Volts reveal changes in the insertion loss and attenuation performance of the filter. Typical surge currents through ground circuits may degrade filter performance and conducted emissions through the filter may increase. These and other emissions and immunity performance criteria may need to be verified through repetitive surge testing. With repetitive surge testing, which is a critical factor in determining the robustness of the filter and which is more representative of the real electrical environment in a facility, some significant changes in insertion loss and attenuation performance may occur. A defined number of surge events, up to 1000 events, including a pause time to allow for component cooling will be injected into the noise filters. The degree of insertion loss and attenuation performance change is determined by subjecting the surged filters to Task 1 to determine if the change is acceptable.

After the 6-kV, 3-kA 8/20-µsec Combination Wave test, the 10-kA 8/20-µsec Combination Wave current pulse will be injected into selected filters. If the filters survive this test without physical damage, the degree of insertion loss and attenuation performance change will be determined by subjecting the surged filters to Task 1 to determine if the change is acceptable.

Results of Test 8
Figure 7 illustrates the filter insertion loss performance for the 5-Amp Ground-Nite filter before and after the 8 x 20 µsec Combination Wave voltage surge test. The “Final Sweep” curve below is the filter performance after the filter was subjected to initial surges starting at 1 kV up to 6 kV in increments of 1 kV, and then subjected to 1000 sequential surges at 6 kV. As one can see from Figure 4, the performance of the filter was not affected by the voltage surges.

Applications
Any EMC investigator knows the difficulties encountered when trying to resolve EMI problems. Moreover, no single EMI solution can be viewed as a “solve-all” technology. Although applying ground noise filters to electrical systems and end-use equipment is a promising type of solution for EMI problems involving the ground system, careful consideration should be exercised to ensure that the solution(s) selected is cost-effective and does not allow other EMI-related problems to surface. The manufacturer of the filter has already solved a number of ground-related EMI problems with end-use equipment using the ground noise filters. These applications range from semiconductor fabrication equipment to automatic parking garage gating systems. Such applications may be more cost-effective than traditional solution approaches. Additional applications of these filters are being examined through field studies in the areas of industrial systems (e.g., adjustable speed drives and programmable logic controllers), building control systems, data center equipment, and medical systems and equipment.

Conclusion
It is a known fact that the electromagnetic environment is becoming more energetic and is cluttered with radiated and conducted emissions from various types of electrical systems and end-use equipment. An appreciable number of EMI problems involve emissions that circulate through electrical conductors within a facility and within equipment. Utilization of ground noise filters to reduce or to prevent such circulation and resolve EMI problems holds promise, especially compared the performance of power line filters.

Acknowledgements
EPRI would like to acknowledge Hideyuki Yamanaka of EMC, Inc., the manufacturer of the ground noise filters tested and Tom Inukai of Sun-Wa Technos America for their contributions and support in conducting this important EMC research.

Bibliography

1. ANSI C62.41.2, American National Standard – IEEE Recommended Practice on Characterization of Surges in Low-Voltage (1000 V and Less) AC Power Circuits. (2003).
2. P.F. Keebler and K.O. Phipps, “Power Quality Effects on the Reliability and Susceptibility of EMI Filters,” Interference Technology EMC Directory and Design Guide, 2007.

Philip F. Keebler has conducted System Compatibility Research on personal computers, lighting, medical equipment, and Internet data center equipment. The lighting tasks were associated with characterizing electronic fluorescent and magnetic HID ballasts, electronic fluorescent and HID ballast interference, electronic fluorescent and HID ballasts failures, electronic fluorescent and HID lamp failures. He has drafted test protocols and performance criteria for SCRP tasks relating to PQ and EMC. Mr. Keebler also manages the Electromagnetic Compatibility (EMC) Group at EPRI where EMC site surveys are conducted, end-use devices are tested for EMC, EMC audits are conducted, and solutions to electromagnetic interference (EMI) problems are identified. He has completed his service as editor developing a new EMC standard for power-line filters, IEEE 1560.

Kermit O. Phipps is a NARTE Certified EMC engineer and conducts tests and evaluations of equipment performance in accordance with standards of ANSI/IEEE, IEC, U.S. Military, and UL, as well as with the EPRI System Compatibility Test Protocols for EPRI. He served in the U.S. Air Force as a manual electronic warfare and component specialist, resolving hardware and software problems. He conducts research on surge protection, power-line filters, shielding effectiveness, and electromagnetic interference. Mr. Phipps is the author and co-author of test plans, protocols, and research papers presented at international power quality and EMC conferences. He has conducted a number of power quality and EMC training sessions and field investigations. Most recently, he has completed his voluntary work as chairman in developing a new EMC standard for power-line filters, IEEE 1560.