An electromagnetic interference analysis of processing environment

An electromagnetic interference analysis of processing environment
Calhoun: The NPS Institutional Archive
Theses and Dissertations
Thesis Collection
2002-12
An electromagnetic interference analysis of
uninterruptible power supply systems in a data
processing environment
Beran, Edward W.
Monterey, Calif. Naval Postgraduate School
http://hdl.handle.net/10945/4314
NAVAL POSTGRADUATE SCHOOL
Monterey, California
THESIS
AN ELECTROMAGNETIC INTERFERENCE ANALYSIS OF
UNINTERRUPTIBLE POWER SUPPLY SYSTEMS IN A DATA
PROCESSING ENVIRONMENT
by
Edward W. Beran
December 2002
Approved for public release; distribution is Richard
unlimited.W. Adler
Thesis Advisor:
Second Reader:
Wilbur R. Vincent
Approved for public release; distribution is unlimited.
THIS PAGE INTENTIONALLY LEFT BLANK
REPORT DOCUMENTATION PAGE
Form Approved
OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction, searching
existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this
burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington headquarters Services,
Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management
and Budget, Paperwork Reduction Project (0704-0188) Washington DC 20503.
1. AGENCY USE ONLY (Leave blank)
2. REPORT DATE
3. REPORT TYPE AND DATES COVERED
December 2002
Master’s Thesis
4. TITLE AND SUBTITLE
5. FUNDING NUMBERS
An Electromagnetic Interference Analysis of Uninterruptible Power Supply Systems in a
Data Processing Environment
6. AUTHOR(S)
Beran, Edward W.
8. PERFORMING ORGANIZATION
REPORT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
Naval Postgraduate School
Monterey, CA 93943-5000
9.
SPONSORING / MONITORING AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSORING / MONITORING
11. SUPPLEMENTARY NOTES
The views expressed in this thesis are those of the author and do not reflect the official policy or position of the Department of
Defense or the U.S. Government.
12a. DISTRIBUTION / AVAILABILITY STATEMENT
12b. DISTRIBUTION CODE
Approved for public release; distribution is unlimited.
ABSTRACT (maximum 200 words)
The levels of Electromagnetic Interference (EMI) generated by two standard models of Uninterruptible Power Supplies
(UPS) were examined. Conducted current measurements were made on all conductors exiting and entering two standard UPS
units between the frequency range of 60-Hz up to 50 MHz. EMI reduction actions were undertaken on both units, and the
reduction in EMI current resulting from these actions was determined. The before and after mitigation results were compared
with EMI limits suggested by available specifications, standards, and other related documents.
The results show that a significant reduction in the level of EMI can be achieved in low-to-modest size UPSs using
inexpensive, standard, and commercially available filters, provided the filters are installed in an effective manner. The
reduction of EMI to harmless levels at radio-receiving and data-processing sites is shown to be feasible.
14. SUBJECT TERMS
15. NUMBER OF
PAGES 126
Conducted EMI, Spectrum Analyzer, Analog Filter, Power Line Filters, UPS
16. PRICE CODE
17. SECURITY
CLASSIFICATION OF
REPORT
18. SECURITY CLASSIFICATION OF
THIS PAGE
Unclassified
19. SECURITY CLASSIFICATION OF
ABSTRACT
Unclassified
Unclassified
NSN 7540-01-280-5500
ANSI Std. 239-18
20. LIMITATION OF
ABSTRACT
Unlimited
Standard Form 298 Rev. 2-89) Prescribed by
i
THIS PAGE INTENTIONALLY LEFT BLANK
ii
Approved for public release; distribution is unlimited.
AN ELECTROMAGNETIC INTERFERENCE ANALYSIS OF
UNINTERRUPTIBLE POWER SUPPLY SYSTEMS
IN A DATA PROCESSING ENVIRONMENT
Edward W. Beran
Civilian, Fort Meade, Maryland
BSEE, University of Maryland, 1973
Submitted in partial fulfillment of the
requirements for the degree of
MASTER OF SCIENCE IN ELECTRICAL ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL
December 2002
Author:
Edward W. Beran
Approved by:
Richard W. Adler
Thesis Advisor
Wilbur R. Vincent
Second Reader
Donald P. Lyons
Fort Meade Advisor
John P. Powers
Chairman, Department of Electrical and Computer Engineering
iii
THIS PAGE INTENTIONALLY LEFT BLANK
iv
ABSTRACT
The levels of Electromagnetic Interference (EMI) generated by two standard
models of Uninterruptible Power Supplies (UPS) were examined. Conducted current
measurements were made on all conductors exiting and entering two standard UPS units
between the frequency range of 60-Hz up to 50 MHz. EMI reduction actions were
undertaken on both units, and the reduction in EMI current resulting from these actions
was determined. The before and after mitigation results were compared with EMI limits
suggested by available specifications, standards, and other related documents.
The results show that a significant reduction in the level of EMI can be achieved
in low-to-modest size UPSs using inexpensive, standard, and commercially available
filters, provided the filters are installed in an effective manner. The reduction of EMI to
harmless levels at radio-receiving and data-processing sites is shown to be feasible.
v
THIS PAGE INTENTIONALLY LEFT BLANK
vi
TABLE OF CONTENTS
I.
INTRODUCTION.............................................................................................. 1
II.
ELECTROMAGNETIC INTERFERENCE................................................ 3
A. BACKGROUND............................................................................................... 3
B. TYPICAL SOURCES OF EMI......................................................................... 4
C. VICTIM DEVICES AND EQUIPMENT ......................................................... 4
D. EMI STANDARDS........................................................................................... 5
E. EMI MITIGATION........................................................................................... 6
III.
EMI MEASUREMENTS ................................................................................ 9
A. TEST SETUP DISCUSSION............................................................................ 9
B. MEASUREMENT CONFIGURATION......................................................... 12
C. TEST LOCATIONS........................................................................................ 13
D. UNMODIFIED UPS RESULTS..................................................................... 14
1. UPS Models Examined ....................................................................... 14
2. Ambient EMI Current ......................................................................... 15
3. EMI Current, Input Conductors .......................................................... 17
4. EMI Current, Output Conductors........................................................ 25
E. MODIFIED UPS RESULTS........................................................................... 33
1. Modifications ...................................................................................... 33
2. Modified UPS Results......................................................................... 34
F. PROPOSED UPS EQUIPMENT SPECIFICATIONS ................................... 42
vii
IV. SUMMARY AND RECOMMENDATIONS .............................................. 45
A. SUMMARY .................................................................................................... 45
B. RECOMMENDATIONS ................................................................................ 46
APPENDIX A. SOURCES OF EMI AND NOISE........................................... 49
APPENDIX B. EMI RELATED TECHNICAL STANDARDS ................... 53
A. INTRODUCTION........................................................................................... 53
B. DISCUSSION ................................................................................................ 55
C. DOCUMENTS ............................................................................................... 56
1. American National Standards Institute (ANSI) 63.12 ........................ 56
2. FCC Regulations: Class B Conducted Limits ..................................... 57
3. MIL-STD-461E (CE-102) Conducted Emissions ............................... 58
4. US NAVY SNEP Documentation....................................................... 58
APPENDIX C. DESCRIPTION OF UPS SYSTEMS ....................................... 61
A. GENERAL ...................................................................................................... 61
B. UPS COMPONENTS ..................................................................................... 62
C. UPS STANDARD........................................................................................... 62
D. UPS INVERTER SECTION........................................................................... 63
E. UPS HAZARDOUS WARNING LABELS ................................................... 64
APPENDIX D. CALIBRATION ........................................................................... 67
A. GENERAL ...................................................................................................... 67
B. TIME-AXIS CALIBRATION OF THE 3-AXIS DISPLAY.......................... 67
C. AMPLITUDE CALIBRATION...................................................................... 66
D. BANDWIDTH CALIBRATION .................................................................... 69
viii
E. PROBE CALIBRATION................................................................................ 70
APPENDIX E. POWER LINE FILTERS............................................................. 73
A. INTRODUCTION........................................................................................... 73
B. FILTER CHARACTERISTICS ..................................................................... 76
1. General .................................................................................................. 76
2. Insertion Loss ........................................................................................ 76
C. FILTER STANDARDS .................................................................................. 77
D. INPUT FILTER............................................................................................... 78
E. OUTPUT FILTER........................................................................................... 78
F. INSTALLATION CRITERIA ........................................................................ 78
G. FILTER MANUFACTURERS ....................................................................... 81
APPENDIX F. PROPOSED SPECIFICATION ................................................. 83
LIST OF REFERENCES ......................................................................................... 101
BIBLIOGRAPHY ...................................................................................................... 103
INITIAL DISTRIBUTION LIST ........................................................................... 105
ix
THIS PAGE INTENTIONALLY LEFT BLANK
x
ACKNOWLEDGEMENTS
The author would like to acknowledge all financial, personal and technical
support provided for the accomplishment of this project.
Funding was provided for two trips to NPS by the Fort Meade National
Cryptologic School (NCS). In addition, financial assistance was given from the NPS
Signal-to-Noise Enhancement Program (SNEP) Laboratory for all lab supplies including
camera film, lab equipment, and drafting support.
Personal thanks to Professor Vicente Garcia for allowing two visits to NPS, thus
enabling completion of performance data. Additionally, gratitude for his ever-present
enthusiasm and patience given to a student in dire need of guidance during an extremely
taxing work - school adjustment period. Professor Garcia kept this student from giving in
to despair during the initial period where reality of the enormous effort required for
upcoming tasks was actually recognized.
Most importantly, the author offers profuse gratitude to Professors Richard Adler
and Wilbur (Ray) Vincent for their expert guidance along with unending patience to a
distraught student during the research and writing of this thesis. To Professor Adler:
thanks for the overall guidance and taking on this student (however taxing he may have
been). Professor Adler, this author recognizes this assistance was an additional burden to
a gentleman with a constantly hectic schedule. To Professor Vincent: thank you for the
intermediate guidance, constant support, and unimaginable, unending patience. Professor
Vincent guided this student around many grammatical and graphical problems with
unbelievable grace and kindness.
Special recognition goes to Mr. Andrew Parker for his generous assistance in all
lab procedures (especially on the student's TDY weekends).
Finally, the author wishes to offer appreciation to Mr. Donald Lyons (the Fort
Meade advisor) for his constructive comments for the completion of the thesis.
xi
THIS PAGE INTENTIONALLY LEFT BLANK
xii
EXECUTIVE SUMMARY
Electromagnetic interference (EMI) generated by small- and medium-size
Uninterruptible Power Supplies (UPS) is investigated in this thesis, and a solution to
minimize the impact of UPS-generated EMI at radio-receiving sites and data-processing
facilities is described. The investigation was instigated by field reports of the Navy’s
Signal-To-Noise Enhancement program (SNEP) teams that have documented numerous
cases of harmful interference to the reception of radio signals at receiving sites. In
addition, cases of the corruption of digital signals in data-processing systems from UPS
EMI have been noted. Since the use of small and medium UPS to prevent power
interruptions to critical equipment is commonplace, a practical and effective solution to
the EMI problem is considered essential.
An introduction to power quality problems and EMI is presented in Section II of
the thesis. EMI can be defined as any electrical signal that adversely affects the operation
of electrical, electronic, or communications equipment. EMI can be generated and
radiated from many devices. The radiated electromagnetic fields can be intercepted by a
victim device and interfere with its operation. Additionally, EMI can also be conducted
from a source to a victim over conductors such as power wires, grounds, cable shields,
and other conducting objects. Descriptions of typical sources of EMI, victim devices and
equipment, and EMI standards are presented. The final discussion presented in this
section concerns EMI mitigation.
In the EMI measurement section of the report, the test setup is explained in detail.
The measurement configuration and test locations are discussed. The UPS measurement
section was divided into two sub-sections, the unmodified (standard, no filters installed)
UPS results and the modified (with EMI filters) UPS results. The modified UPS results
show that EMI currents have been reduced significantly with the use of EMI filters on the
input and output conductors of the UPS. The final subsection is a proposed UPS
equipment specification.
xiii
Appendix A provides a “Table of Noise Sources” obtained from the United States
Signals Intelligence Directive. The noise sources or emitters shown in Table A.1 of the
Appendix are intended only for general guidance to planners and operational units since
it is impossible to list all possible sources. While only a limited number of sources
existed in past years, the recent introduction of new devices into Department of Defense
facilities (especially digital power-control devices) has introduced many new kinds of
sources. Each situation involving electrical and radio-noise problems must be evaluated
on a case-by-case basis to identify and mitigate any adverse input on facility operations
from each individual source. Existing noise sources in Department of Defense facilities
cannot always be summarily removed because of overall operational considerations, but
each source can and should be modified or replaced to minimize the deleterious impact of
noise on the operation of electrical and electronic systems. Every effort must be made to
protect against ongoing mission performance degradation created by EMI. Such actions
will improve the performance of radio and data-processing systems and often results in
the improved efficiency of the devices causing noise.
In Appendix B, a document search was performed to find existing EMI-related
documents that identify EMI current criteria standards. This investigation of various
standards was accomplished to determine the latest interference criteria at the time of the
writing of this thesis. Based on several document searches, it was decided to use the
SNEP interference criteria for this report.
In Appendix C, a description of UPS systems is presented. Two typical office
data-processor-type UPS systems were used in the analysis of this document. The
components of UPS (rectifier or charging unit, inverter, and battery bank) were briefly
discussed. UPS standards and warning signs were next discussed.
In Appendix D, the test calibration methods are explained. The descriptions cover
time-axis calibration of the 3-axis display test equipment, amplitude calibration,
bandwidth calibration, and probe calibration techniques.
In Appendix E, an introduction to EMI filter theory is next presented.
Descriptions of t- and pi-filters are discussed. For the purposes of this thesis, the
discussion will be limited to the low-pass type, which allows alternating electrical power
xiv
to be provided to an UPS while attenuating higher-frequency EMI being conducted out of
the UPS. A brief discussion of filter characteristics, insertion loss, standards, and finally
installation criteria are given.
In Appendix F, the engineering specification for a single-phase UPS system is
given. Section 16611 from the MASTERSPEC DRAWING COORDINATION
document issued by the American Institute of Architects (AIA) provides an overall
specification for an Uninterruptible Power Supply (UPS). Suggested changes in this
specification are listed in bold. These changes are directed at the use of UPS systems in
radio-receiving and sensitive data-processing facilities. Of particular concern is the
elimination of harmful levels of UPS generated EMI on radio signal reception and on the
operation of sensitive data-processing systems.
In the summary section of the thesis, the author concluded that EMI can degrade
the operation and performance of many kinds of electronic devices and radio receiving
systems. The increasing use of computers, digital-data-processing devices, and powercontrol devices in such facilities (including UPS systems, switching power supplies, and
motor controllers based on solid-state switching techniques) has resulted in many cases of
EMI problems. In UPS systems, EMI is generated by the rectifier and inverter sections,
which often produce noise up to and sometimes higher than 50 MHz. Also, another
intermittent source of impulsive noise is load-current changes (load switching), that
create voltage and current impulses and electrical noise. The EMI generated by two
unmodified commercial models of UPS were examined. The units were modified in
accordance with integrated Barrier, Filter, Ground techniques and again tested.
Significant reductions in conducted EMI current was achieved on all conductors
penetrating the UPS case including all power and ground conductors. The modifications
resulted in the test units meeting all known conducted EMI current limits including those
of MILSTD-461 and the suggested limits provided by the US Navy SNEP team.
The MASTER DRAWING COORDINATION document issued by the American
Institute of Architects (AIA) is often used as a guide for the procurement and installation
of UPS systems. This document, in its original form, does not consider the impact of
xv
UPS-generated EMI on other systems. Based on the results of this investigation,
suggested changes to the AIA document are provided in Appendix F.
In the recommendation section of the thesis, the author offered the following
suggestions:
•
To follow the Naval Security Group (NSG) recommendation for the use of
UPS (see page 46): “Only mission-essential equipment not able to tolerate
even momentary power disturbances shall be connected to UPS power.” This
recommendation is fully supported, and the DOD will be best served by
following the NSG recommendation.
•
In addition, to the above general recommendation, additional precautions need
to be taken to ensure that UPS-generated EMI does not degrade the operation
of other radio and electronic systems. This is especially the case in HF, VHF,
and UHF radio-receiving sites where UPS-generated EMI is often found at the
input terminals of the radio receivers and in sensitive data-processing facilities
where UPS-generated EMI is often found on data-cable shields and grounds.
•
When a determination is made that an UPS is required, or an UPS already
exists, specific guidelines and critical aspects for consideration both prior to
purchase and after the product has been obtained or installed should be
followed.
•
Although an UPS provides excellent protection for sensitive equipment from
many kinds of power faults, their generation of low- and high-frequency EMI
must be considered. At times, interference may exist after the UPS has been
installed. During these occasions, the procedure for identifying and
documenting harmful levels of EMI becomes somewhat more complicated.
•
In cases where UPS noise may be a factor degrading the performance of
electronic and radio devices or systems, it is recommended that UPS be
procured and installed in accordance with the additions to “The MASTER
DRAWING COORDINATION document issued by the American Institute of
Architects (AIA)”. These additions and changes are provided in Appendix F.
xvi
These changes apply to small and medium power UPS systems (up to about
50 kVA) where standard filters are available as COTS items.
•
Finally, it is recommended that future projects be undertaken to better
understand methods to control EMI caused by a medium-sized (i.e., 50 to 250
kVA) UPS and large sizes (i.e., 250 kVA and higher). A project is required to
develop methods and filter configurations to lower conducted and radiated
EMI from medium and high power UPS to acceptable limits.
xvii
THIS PAGE INTENTIONALLY LEFT BLANK
xviii
I.
INTRODUCTION
Field experience has shown that electromagnetic interference (EMI) generated by
Uninterruptible Power Supplies (UPS) and other similar solid-state-switching devices can
result in the degradation of signal reception at radio-receiving sites and the contamination
of data at data-processing facilities. The need for UPS systems to improve power
reliability for critical devices at such sites is recognized, and this indicates the need for
devising effective EMI mitigation actions. This matter is examined in this thesis.
A review of available EMI standards, specifications, and other related documents
was undertaken. This included Department of Defense standards and documents as well
as standards and documents of other agencies such as the Federal Communications
Commission (FCC) and the Institute of Electrical and Electronic Engineers (IEEE).
Although no standard was found that fully covered the UPS/EMI issue, all addressed the
topic to some extent.
EMI tests were conducted on two standard UPS units over the frequency range of a
few kHz and as high as 50 MHz. EMI suppression actions were then undertaken on both
units and the amount of EMI reduction was measured. The method of EMI suppression is
described. The results before and after suppression actions were compared to conductedcurrent limits suggested by available standards, specifications, and documents.
The Appendices provide supplemental information on topics associated with
electrical noise and used for research in this thesis. The topics include electrical noise
sources, EMI standards, UPS system descriptions, test calibrations, EMI filter
characteristics, and UPS technical equipment specifications.
1
THIS PAGE INTENTIONALLY LEFT BLANK
2
II.
ELECTROMAGNETIC INTERFERENCE
Factors important in the evaluation of electromagnetic interference from an
Uninterruptible Power Supply are presented in this chapter.
A.
BACKGROUND
The operation of sensitive electronic components and systems can typically prove
to be adversely affected by a number of power-quality factors such as EMI. An
understanding of these factors is required for investigation of, and solutions to EMI
occurrences. Power quality issues are specifically described by the following general
categories:
♦ Transients - Disturbances with high-speed voltage or current changes.
Transients are sometimes described as spikes, impulses, and surges.
♦ Momentary interruptions - Referring to a loss of voltage for periods of less
than a cycle to several cycles.
♦ Sags and swells - Variations in voltage or current, ranging from one half
second to several seconds. Sags refer to reductions in voltage or current while
swells refer to increases in voltage or current.
♦ Under voltage or over voltage - Described as sags or swells continuing for
more than several seconds or at times lasting for longer than a period of hours.
♦ Harmonic distortion - Occurrences where the waveshape of voltage or current
is not sinusoidal.
EMI is one additional factor often affecting the operation of sensitive electrical and
electronic equipment. EMI can be defined as any electrical signal that adversely affects
the operation of electrical, electronic, or communications equipment. EMI can be
generated and radiated from many devices (or from the conductors associated with such
devices). The radiated electromagnetic fields can be intercepted by a victim device and
interfere with its operation. Additionally, EMI can also be conducted from a source to a
victim over conductors such as power wires, grounds, cable shields, and other conducting
objects. Both forms of transmission regarding EMI from a source to its victim must be
considered when investigating EMI problems.
3
Two terms are often used to describe conducted EMI; they are known as commonmode and differential-mode. Common-mode EMI is the interference measured between a
conductor (or a pair of conductors carrying a desired signal) and a reference ground.
Differential-mode EMI is interference measured between two conductors carrying desired
signals. Conducted EMI interference can be measured and compared with either voltage
or current established criteria.
B.
TYPICAL SOURCES OF EMI
EMI is often generated by a variety of electrical and electronic devices and
equipment. Typical types of equipment that can create EMI are:
♦ Uninterruptible Power Supplies
♦ Telecommunication Equipment
♦ Personal Computers and Peripheral Equipment
♦ Switching Power Supplies
♦ Variable Frequency Induction Motor Drives
♦ Battery Chargers
♦ Electronic Dimmers
♦ Electronic Fluorescent Light Ballasts
♦ RF Stabilized Arc Welders
♦ Some Medical Equipment
♦ Power Conversion Devices Based on Switching Techniques
Additional sources [Reference 1] of EMI are provided in Appendix A and derived
from the United States Signals Intelligence Directive.
C.
VICTIM DEVICES AND EQUIPMENT
Most electronic devices can be adversely affected by EMI including items in the
list of sources provided in Section B, but data-processing equipment, communications
4
devices, and radio receivers located in remote facilities are the items of highest concern in
this thesis.
The operating speeds of digital devices and equipment continue to increase,
making them more susceptible to high-frequency EMI. In addition, the operating voltage
of many solid-state devices will continue to decrease while conserving power, thereby
creating susceptibility for malfunction due to transients and other forms of EMI. Of
primary concern are these two trends in device and equipment design, combined with the
increased usage of various devices causing EMI.
D.
EMI STANDARDS
A large number of publications, references, handbooks, and standards exist
addressing the topic of EMI. The pertinent handbooks and standards are listed and
reviewed in Appendix B. A number of government handbooks and standards provide
guidance concerning EMI. However, the enormous variety of electronic and electrical
devices, ongoing introduction of new devices and equipment, and the long production
times for revision of standards and handbooks causes difficulty in covering total aspects of
EMI. In addition, there is an increasing tendency to rely on commercial-off-the-shelf
(COTS) equipment and associated standards for the procurement and installation of
electrical and electronic equipment.
Much of the equipment in today’s data-processing centers and radio-receiving
facilities is purchased in accordance with COTS requirements. For example, the UPS
systems of primary concern within this thesis are almost always procured as COTS
equipment, where the primary standard is the Class A or Class B requirements of Part 15
of the Federal Communications Commission. The Class A requirement is directed at
providing equipment to be installed and used in industrial facilities. A far more stringent
Class B requirement is directed at the provision of equipment for use in residences.
5
E.
EMI MITIGATION
A commonly used approach to EMI mitigation, better and bigger grounds (or
ground impedance), was used with much success in past decades when cases of highfrequency EMI current flowing on grounds, power conductors, cable shields and other
conductors at a facility was rare. Grounds provided a means to establish a voltage
reference for all equipment in a facility. The more recent introduction of solid-state
switching devices and other digital devices into electronic and power-control equipment
greatly increased the amount and levels of both low-frequency and high-frequency EMI
current and voltages. This occurs on grounds, power conductors, cable shields, equipment
cases, and other conductors. In addition, the lengths of conductors associated with EMI
sources, victim equipment, and the paths between the two become electrically long at
higher-frequency components of EMI. Multiple wavelength paths now introduce standing
waves of EMI voltage and current on these paths, which further complicates coupling
mechanisms between source and victim. These factors significantly reduce the
effectiveness of the grounding approach. Electrical lengths of grounds, power conductors,
cable shields, and other conductors are now a key factor in the EMI problem rather than a
part of the solution. Transmission line and antenna theory is now an integral part of an
EMI problem.
Moreover, efficient near-zone coupling mechanisms (both inductive and
capacitive) allow high-frequency EMI current and voltage from a conductor to be coupled
onto other closely-spaced conductors at a facility. The near-zone coupling mechanisms
and direct conduction of EMI current from a source to another location over multiple
paths allows EMI current and voltage to spread far beyond its source and to seek paths of
entry into victim equipment. This can result in a large number of paths for the flow of
EMI current making the description and modeling of a source, its paths to a victim, and
the susceptibility of a victim very difficult and often impractical to identify.
A description of UPS systems is given in Appendix C. A simplistic but effective
approach to the mitigation of EMI from an UPS (or any other source) was taken during
this effort. The approach was discovered in the early era of radio and telegraph
communications. It consists of merely preventing EMI current from flowing on any
6
conductor outside the case or housing of a device generating EMI along with the
recognition that grounds are now a part of the EMI problem rather than a solution.
This approach is described in a number of unclassified publications by Nanevich,
Vance, and Graf of SRI International [References 2 through 5]. These publications
describe an EMI control technique they call the “Topological Control of EMI.” It consists
of the use of shielding around a source or victim, applying meaningful bandwidth control
on all conductors entering and exiting the shield, the termination of internal cable shields
and grounds on the inside surface of the shield, and the termination of external cable
shields and facility ground conductors on the outside surface of the shield. This concept
has been extended and used by USN Signal-To-Noise-Enhancement Program teams.
These teams applied usage of the electromagnetic barrier, filter, ground (BFG) technique
[Reference 6] to control EMI from sources, or lowering susceptibility of victims. The
BFG technique was applied to the test UPS systems examined in this thesis.
The general background and information needed for the control of harmful levels
of electrical and electromagnetic interference from UPS systems has been presented
above. This background has been stated in terms used by and encountered by radioreceiving site and sensitive data-processing site personnel.
7
THIS PAGE INTENTIONALLY LEFT BLANK
8
III
EMI MEASUREMENTS
This Chapter describes the measurement setup and presents the results of
measurements on one Interruptible Power Supply of a type frequently found in field sites
and data-processing facilities.
A.
TEST SETUP DISCUSSION
The test equipment utilized in this study is identical to the setup used by the NPS
Signal-Enhancement Laboratory and for the Naval Security Group (NSG) Signal-to-Noise
Enhancement Program (SNEP). This equipment has been used for many laboratory and
field measurement programs. All tests and measurements provided in this document are
based on conducted emissions from standard COTS UPS systems. The measurements
provide values of conducted EMI current over broad bandwidths before and after the UPS
systems were modified where the modifications were made to reduce conducted EMI
current to harmless levels. Figure 3.1 shows a photograph of the primary items of
instrumentation in this test setup.
Figure 3.1: The SNEP Test Setup
9
Figure 3.1 shows a low-frequency spectrum analyzer in the lower right part of the
view, a high-frequency spectrum analyzer in the lower left part of the view, the timehistory display on top of the high-frequency spectrum analyzer and line amplifiers on top
of the time-history display. The current probes used to measure conducted EMI current
are not shown.
The instrumentation setup provides a capability to measure low-frequency EMI
current over the frequency range of 0 to 100 kHz in addition to high-frequency EMI
current over the frequency range of 50 kHz up to 100 MHz. Both capabilities were used
to produce the examples of data shown in this document. Figure 3.2 shows a block
diagram of the instrumentation used for the low-frequency measurements.
CURRENT
PROBE
LINE
AMPLIFIER
SPECTRUM
ANALYZER
OSCILLOSCOPE
CAMERA
Figure 3.2: Block Diagram of the Low-Frequency Test Setup
A clamp-on Tektronix Model CT-4/P6021 current probe was provided to measure
low-frequency EMI current on power and ground conductors. The frequency range of the
probe was much larger than the frequency range of any measurement, but the full
frequency range could not be utilized. The maximum amplitude of the low- or highfrequency components of EMI current and the lowest levels of EMI current occurring at
higher frequencies exceeded the dynamic range of the instrumentation. This prevented
simultaneous measurement of low- and high-frequency currents. A Hewlett-Packard
Model 3561 spectrum analyzer was used to measure the amplitude of spectral components
of low-frequency EMI. A WRV Model A-102 line amplifier was provided to measure
10
very low levels of EMI current. The line amplifier was seldom used, since most
components of EMI current were high enough to be examined directly by the spectrum
analyzer. A Tektronix Model C-5C oscilloscope camera was used to photographically
record examples of EMI current.
A second set of instrumentation was used to measure high-frequency EMI current.
Figure 3.3 shows a block diagram of this instrumentation.
CURRENT
PROBE
LINE
AMPLIFIER
SPECTRUM
ANALYZER
3-AXIS
DISPLAY
OSCILLOSCOPE
CAMERA
Figure 3.3: Block Diagram of High-Frequency Test Setup
A clamp-on Fischer Model F-70 current probe was used to measure EMI current
over the frequency range of 100 kHz through 100 MHz. A calibration curve was provided
to extend the frequency range of the probe down to 50 kHz. This provided some overlap
in the frequency range of the low- and high-frequency instrumentation systems. A high
dynamic-range line amplifier providing a gain of 20 dB was used to increase the
measurement range of the setup to low values. A standard Hewlett Packard Model-141
spectrum analyzer was used to examine the spectral and temporal structure of conducted
EMI current. An ELF Engineering 3-axis display was used to provide a time-history view
of variations in the level and frequency of spectral components of EMI current and to
allow examination of the temporal structure of impulsive components of the current. A
Tektronix Model C-5C oscilloscope camera was used to photographically record examples
of conducted EMI current.
11
Figure 3.4 shows the format of the time-history presentation. A small
identification chart was added to each example of data presented at the lower left part of
the data. The information in this chart is:
Line 1 Date of the measurement in yymmdd format, local time
Line 2
Site ID, UPS ID
Line 3
Conductor ID, UPS Load Information
Line 4 Center Frequency, Scan Width, IF Bandwidth, Scan Time
Line 5
Line Amplifier Gain, RF Attenuation, IF Gain
Line 6
Additional Information
Figure 3.4: Format of the 3-Axis Display
B.
MEASUREMENT CONFIGURATION
Figure 3.5 shows a photograph of the UPS under test. APPENDIX C presents a
brief system description of the UPS systems. The UPS under-test is mounted on a small
pallet and the input and output power conductors are shown in the photograph. Power
cord adapters were used to allow for measurement of current on individual conductors of
the power cord or the common-mode current on all conductors. A Fischer Model F-70
current probe is shown in the lower part of the view along with a section of cardboard
used to insulate the probe from the cement floor of the test location.
12
Figure 3.5: UPS under Test
Combinations of tests were performed for purposes of this study. First, a calibration test was performed. A no-load test was then completed to examine the ambient current flowing on the power conductors. This was followed by an on-off test to compare the
ambient levels of current with the UPS operating levels of current. Finally, a test was
conducted with the UPS operating with a resistive load. These sequences of tests were
performed on each UPS, first with the COTS configuration of the UPS, and second with a
filter added to the UPS input and output conductors.
The detailed test procedures are provided in Appendix D. The calibration curves
for the instrumentation are also provided in Appendix D.
C.
TEST LOCATIONS
Two test locations in California were used for measurement. The first location
was at a small, suburban, radio-receiving site located in Los Altos Hills, CA. This was a
noise-quiet radio-receiving site used for special signal-reception tests. The second
location was at the Signal Enhancement Laboratory in Spanagel Hall of the Naval
Postgraduate School in Monterey, CA.
13
While the first site was considered to be a low-noise radio-receiving site, it was
located fairly close to several high-power medium-frequency broadcast stations in the San
Francisco region. Additionally, it was fed from overhead power lines although sources of
noise from hardware on the power-line had been eliminated.
The second site was fed from underground power lines, but the facility contained
laboratories with electrical and electronic equipment. Included were a large number of
power-conversion devices, several motor controllers, a large number and variety of
computers, data-processing equipment, and other sources. These sources impressed
significant levels of ambient EMI current onto the power conductors.
D.
UNMODIFIED UPS RESULTS
1.
UPS Models Examined
A standard Model Ferrups FE/QFE 1.4 kVA UPS manufactured by the Best
Corporation of Necedah WI was obtained for the investigation of EMI. In addition, a
Model Ferrups FE/QFE 2.1 kVA was also available, as well as models from other
manufacturers. All UPS units examined were new and in excellent operating condition.
Past experience with UPS units installed in field sites suggested the Model Ferrups
FE/QFE 1.4 kVA generated less, but unknown amounts of EMI, compared to other
models from Best or from many other manufacturers. Thus, EMI levels from the
unmodified UPS presented in this thesis are ostensibly lower than many other similar
units. Curiously interesting is that all UPS units investigated during this effort were
advertised as meeting FCC Class A requirements. No UPS could be found meeting the
more stringent FCC Class B requirements, which have stricter guidelines.
Conducted EMI was measured on all conductors penetrating the case of each UPS
during normal operation. No additional conductors or test conductors were allowed to
penetrate the case during the tests. The metal case provided with the UPS was in place
during all measurements to minimize radiated effects. The UPS case was unmodified and
it was installed in accordance with the manufacturer’s instructions.
14
Tests were first made with the UPS as it was obtained from the manufacturer. The
UPS as then modified in accordance with Barrier, Filter, Ground principles [Reference 5]
and re-tested using the same instrumentation.
2.
Ambient EMI Current
During the tests the ambient EMI current on the input power conductors proved
very high at frequencies below 2 MHz, interfering with low-frequency measurements.
Illustrating detrimental effects, EMI current was measured on input power conductors
while the UPS was switched off. Figure 3.6 shows the spectral components of ambient
current on the white, black, and green conductors of the UPS power cord.
The data in Figure 3.6 shows ambient current from two different sources flowing
on all of the input power wires. Signals from nearby broadcast stations induced current
into the overhead distribution line providing electrical power to the test facility. These
signals are represented by the discrete-frequency spectral components shown in each of
the three amplitude-vs.-frequency views. In addition, broadband impulsive noise current
was also flowing on the power wires. The source of this current was later traced to a
recently installed variable-speed drive on an air-conditioning system at a nearby residence
receiving power from the same distribution line as the measurement facility.
Significant variations in the amplitude of the ambient impulsive current are shown
across the 2-MHz band of the data in Figure 3.6, including peaks and nulls. The peaks
and nulls suggest that resonance conditions existed in the power conductors, but this is to
be expected since the conductors are electrically long at the frequency range of the data.
The amplitude of the ambient current can be scaled from the data for any desired
frequency, but a single value of current amplitude cannot be used to provide a meaningful
or complete measure of amplitude. Thus, a meter reading of the EMI current is not
feasible.
15
Figure 3-6: Ambient EMI Current on Input Power Cord Conductors
16
The high ambient current on power conductors is one factor that must be considered during such tests. Impulsive noise from power-conversion equipment based on solidstate switching techniques is now encountered in most facilities. For critical tests it is
necessary to use a diesel-powered generator to avoid contaminated power. The amplitude
of the ambient current from these two sources must be ignored in the evaluation of EMI
from the UPS.
3. EMI Current, Input Conductors
The UPS was then turned on and the battery was allowed to charge to its full level.
When the battery was fully charged, a 300-watt resistive load was switched on. (A
resistive load was used to avoid the harmonics and impulsive EMI current generated by
loads containing nonlinear devices such as switching power supplies, motor controllers,
computers, and other similar devices.) A 50-ft power cord was used between the UPS and
the resistive load to provide an electrically long length of conductor between the UPS and
the load. EMI current was then measured on the three input power conductors and the
three output conductors. The broadband impulsive current from the switching source
was somewhat similar for the black and white conductors in that the amplitude was high
from about 10 kHz up to 1 MHz. It decreased in amplitude and became quite low at 2
MHz. The current on the green-wire ground conductor was low at approximately 100
kHz, reaching peaks in the range of 0.5 to 1 MHz. The current then decreased in
amplitude above 1 MHz and was quite low at 2 MHz.
Broadcast-band signals were found on all three conductors. The presence of the
broadcast-band signals suggest that a site with an underground power feed would be more
suitable since an underground feed would not act as a receiving antenna for such signals.
Figures 3.7 through 3.9 show the low-frequency ambient and UPS-generated EMI
current flowing on the input power conductors when the UPS was operated. Both the
amplitude-vs.-frequency and the time-history views are provided to show the coarse-scale
temporal of impulsive noise. The upper part of the time-history view shows the ambient
17
Figure 3.7: Ambient and UPS EMI Current on the Black Conductor,
Low Frequencies
18
Figure 3.8: Ambient and UPS EMI Current on the White Conductor,
Low Frequencies
19
Figure 3.9: Ambient and UPS EMI Current on the Green Conductor,
Low Frequencies
20
current. The display was temporarily placed into a “freeze-mode” after portraying a short
period of ambient current to allow the operation of the UPS to stabilize. The display was
then unfrozen to show the additional EMI current generated by the UPS. The spectral
structure of both the ambient current and the UPS-generated impulsive current is shown
in the amplitude-vs.-frequency view.
Figure 3.7 shows the low-frequency current flowing on the black input power
conductor over the frequency range of 0 to 2 MHz. In this case, ambient noise was
higher than the UPS generated noise. It reached a sharp peak of about 20 µΑ near 0.5
MHz. It is impossible to read the current levels in the part of the medium-wave broadcast
band containing closely-packed signals with the wide frequency span used to generate the
data in Figure 3.7, and it is necessary to search between broadcast-band signals by
employing a narrow scan width. Figure 3.8 shows ambient and UPS-generated current
on the white input conductor and Figure 3.9 shows the ambient and UPS-generated
current on the green-ground conductor of the power cord.
The high-frequency components at 2 to 12 MHz of EMI current flowing on the
conductors of the input power cord were examined next. In this case, the ambient
currents, with minor exceptions, were well below the UPS-generated levels and in most
cases below the signal-detection sensitivity of the instrumentation.
Figures 3.10 through 3.12 show EMI current levels on the black, white, and green
conductors. The presentation is identical to that used for the low-frequency current
measurements except for the frequency span.
Figure 3.10 shows ambient and EMI current flowing on the black conductor over
the frequency range of 2-to-12 MHz. The current fell below instrumentation sensitivity
at frequencies above 12 MHz for the particular model of UPS being tested, therefore
higher frequency data is not shown. A small amount of ambient current is shown at the
extreme left edge of the frequency range. The broad peak in current centered at about 6
MHz is from the UPS, reaching a level of about 5 µA. A smaller peak in EMI current
was found near 12.5 MHz. The four discrete-frequency spectral components in the upper
21
Figure 3.10: Ambient and UPS EMI Current on the Black Conductor,
High Frequencies
22
Figure 3.11: Ambient and UPS EMI Current on the White Conductor,
High Frequencies
23
Figure 3.12: Ambient and UPS EMI Current on the Green Conductor,
High Frequencies
24
half of the frequency range are HF signals collected by the power wiring. These signals
were present regardless of whether the UPS was switched off, or on, while testing.
Figure 3.11 shows the ambient and EMI current flowing on the white conductor
over the frequency range of 2-to-12 MHz. The ambient current at the low end of the
frequency span was lower than observed with the black conductor, while the broad peak in
UPS-generated EMI current was centered somewhat lower in frequency at about 4 MHz.
The amplitude proved somewhat higher at 11 µA. The higher-frequency peak found on
the black conductor did not appear on the white conductor. The four HF signals did
appear in the data.
Figure 3.12 shows ambient and EMI current flowing on the green-wire ground
conductor of the power cord. A small amount of ambient current was found at the
extreme low end of the frequency scale. The broad spectral peak in EMI current shown
near 4 MHz is wider than for the other conductors, and the measurable EMI current
extended upward to 12 MHz. Narrow peaks and nulls in the amplitude of UPS-generated
EMI current are shown for all conductors suggesting resonant conditions on the power
conductors.
4.
EMI Current, Output Conductors
A similar set of measurements was made on the output conductors running from
the UPS to the resistive load. Figures 3.13 through 3.15 provide levels of EMI current on
the output conductors running from the UPS to the resistive load. The data shows the
current over the frequency range of 0 to 2 MHz although the amplitude readings are
calibrated only over the 0.1 to 2 MHz portion of the data. In this case the ambient current
from impulsive noise was below the levels of UPS-generated current, while discretefrequency signals from the local medium-frequency broadcast-band stations appear in the
data and must be ignored. Figure 3.13 shows EMI current flowing on the black output
conductor. The slanting lines in the time-history view provide a convenient means to
separate the UPS-
25
Figure 3.13: UPS EMI Current on the Black Output Conductor,
Low Frequencies
26
Figure 3.14: UPS EMI Current on the White Output Conductor,
Low Frequencies
27
Figure 3.15: UPS EMI Current on the Green Output Conductor,
Low Frequencies
28
generated noise from the broadcast-band signals. A peak in the UPS-generated noise is
shown at 0.4 MHz. Lower-level peaks in EMI current can be distinguished throughout the
frequency range including a distinct peak near the upper end of the frequency scale at 1.7
to 1.8 MHz.
Figure 3.14 shows EMI current flowing on the white output power conductor. The
results are similar to those obtained from the black conductor.
Figure 3.15 shows EMI current flowing on the green-wire ground conductor
running from the UPS to the resistive load. The view is less cluttered than for the black
and white conductors because less ambient current from broadcast signals is picked up by
this conductor.
Distinctive spectral peaks and nulls in UPS-generated current appear on all
conductors. This is an indicator that resonance conditions exist on all conductors from a
combination of conductor length and the impedance of UPS components in the output
paths. The peaks and nulls in the current makes it impossible to describe the amplitude of
the EMI current with a single number; however, a value of UPS current can be provided at
any selected frequency.
EMI and ambient current flowing on the output conductors over the frequency
range of 2 to 12 MHz was also examined. Figures 3.16 through 3.18 show the UPSgenerated EMI current on the conductors of the output power cord. In this case most of
the ambient current was well below the UPS-generated EMI current and only a small
amount of ambient current appears in the data.
Figure 3.16 shows the current on the black output conductor over the 2- to 12MHz frequency band. Figure 3.17 shows the current on the white output conductor and
Figure 3.18 shows the current on the green-wire ground conductor.
An examination of the data from the three output power conductors shows two
distinct variations in amplitude with frequency. Broad peaks and nulls in the amplitude of
the UPS-generated EMI current were found on all conductors.
29
Figure 3.16: UPS EMI Current on the Black Output Conductor,
High Frequencies
30
Figure 3.17: UPS EMI Current on the White Output Conductor,
High Frequencies
31
Figure 3.18: UPS EMI Current on the Green Output Conductor,
High Frequencies
32
In addition, narrow peaks and nulls also exist in the data. This indicates that two
distinct resonance phenomena exist on the output conductors. These peaks and nulls
prevent providing a single number for the UPS-generated EMI current although a value
can be provided for any specific frequency. Two discrete-frequency peaks in ambient
current appear on the black and white conductors at frequencies near 4 and 8 MHz.
These signals were from nearby HF transmitter facilities; therefore, they can be ignored.
E.
MODIFIED UPS RESULTS
1.
Modifications
Standard COTS filters were added to the input and output power conductors in a
Barrier, Filter, Ground (BFG) Configuration. A description of EMI filters is given in
Appendix E. Figure 3.19 shows the configuration employed to allow 60-Hz current to
flow into and out of the UPS.
The filters provided high impedance to the flow of high-frequency EMI current
through and then out of the UPS on the black and white power conductors at frequencies
above its cutoff frequency. Of special concern was the prevention of UPS-generated EMI
current from escaping the UPS case on the green-wire ground while still maintaining the
required electrical safety requirements of the National Electric Code. If EMI current was
allowed to flow on the green-wire ground, it would be inductively coupled back onto the
black, white, and other conductors thereby negating the effectiveness of the filtering
provided on the black and white conductors.
The flow of current on the green wire was limited to low frequencies by
connecting the external green-wire ground to the outside surface of the UPS case and the
internal green-wire ground to the interior of the UPS Case. This type of connection was
provided by the metal shell of the COTS filters.
Figure 3-20 shows a photograph of the modified UPS. Both the input and the
output filters were added to the back panel of the UPS. Since there was room inside the
UPS case for the filters, the modification was easy to implement.
33
IN T ER IO R
OF
U PS C A SE
FILTER
FILTER
B LA C K
W HITE
IN PU T
POW ER
CONDUCTORS
G RE E N
FILTER
FILTER
B LA C K
W HITE
O U TPU T
POW ER
CONDUCTORS
G RE E N
Figure 3.19: Filter Configuration
Figure 3.20: Rear Panel of the Modified UPS
2.
Modified UPS Results
Measurements on the modified UPS were made at the laboratory facilities of the
Signal Enhancement Laboratory of the Naval Postgraduate School. Unfortunately, the
ambient EMI level at this facility was too high to obtain good low-frequency EMI data.
34
The presence of many other sources of EMI in the building including variable-speed
motor controllers, many computers, switching power supplies, and other digital devices
prevented low-frequency data from being obtained. The ambient interference was
sufficiently high, making the normal reception of radio signals impossible from very low
frequencies to above 30 MHz. In addition, instances of the corruption of data lines with
COTS type data-processing equipment was common in the facility when certain powerconversion devices were operated. Reasonable high-frequency data was obtained in spite
of the high ambient EMI levels. Figures 3.21 through 3.23 shows the high-frequency
EMI current on the UPS input conductors over the 2- to 12-MHz band.
Figure 3.21 shows current flowing on the black input conductor. Some small
peaks in current are shown in the data, but these were all ambient current which was
present whether the UPS was switched on or off during testing.
Figure 3.22 shows the high-frequency EMI current flowing in the white input
conductor. Again, all spectral components shown in this view were caused by other
devices operating in the test facility and not from the UPS under test. No low-level
spectral-component of current could be traced to the UPS.
Figure 3.23 shows the high-frequency EMI current flowing on the input greenwire ground conductor of the UPS power cord. Once again, the current shown in the data
was from other sources, and no spectral component could be traced to the UPS under test.
The lack of high-frequency UPS-generated current on the green-wire ground shows the
effectiveness of the green-wire ground connection employed in the UPS modifications.
This connection provided a conducting path at low-frequencies, meeting the requirements
of the National Electric Code, and it employed the shielding provided by the metal UPS
case to prevent the flow of UPS-generated EMI current to the outside green-wire ground
conductor.
Figures 3.24 through 3.26 shows the high-frequency EMI current flowing on the
output conductors of the modified UPS. Figure 3.24 shows the current flowing on the
35
Figure 3.21: High-frequency EMI Current on the Black Input Conductor,
Modified UPS
36
Figure 3.22: High-frequency EMI Current on the White Input Conductor,
Modified UPS
37
Figure 3.23: High-frequency EMI Current on the Green Input Conductor,
Modified UPS
38
Figure 3.24: High-frequency EMI Current on the Black Output Conductor,
Modified UPS
39
Figure 3.25: High-frequency EMI Current on the White Output Conductor,
Modified UPS
40
Figure 3.26: High-frequency EMI Current on the Green Output Conductor,
Modified UPS
41
black output conductor. The minimum detectable signal for the example shown is about
7 µA peak and 4 µΑ rms using a measurement bandwidth of 100 kHz. Additionally, a
lower-level discrete-frequency signal and/or lower-level narrow band spectral components of noise can be detected with a narrower measurement bandwidth. This was
attempted at measurement bandwidths down to 3 kHz and UPS-generated noise could not
be detected.
Figure 3.25 shows a similar result for the white input power conductor. The
modified UPS again lowered EMI current well below the detection level of the
instrumentation.
Figure 3.26 shows the high-frequency EMI current flowing on the output greenwire ground conductor. In this case ambient current was found, but no UPS-generated
component was detectable. The ambient current resulted from the conducting path
provided by the input green-wire conductor, the external surface of the UPS case, and the
green-wire output ground conductor.
While the configuration with the addition of modifications successfully prevented
UPS-generated current from flowing on input and output conductors, the UPS
modifications could not prevent ambient current generated by other sources from flowing
on facility ground conductors. The ambient current from other sources must be
controlled at each source and the modification of other devices will not correct other
facility problems. This is also a clear indication that ambient current in a facility ground
cannot be eliminated by ground-system modifications.
F.
PROPOSED UPS EQUIPMENT SPECIFICATIONS
Although no manufacturers have EMI filters installed prior to purchase, it is
suggested that engineers, users, or procurers of UPS systems edit UPS equipment
specification (AIA Specification 16611, Appendix F of this thesis) to reflect EMI current
concerns when installing an UPS into a facility containing receiving and sensitive dataprocessing equipment. It is noted that some manufacturers will agree to have EMI filters
installed upon request. During the specification phase of procurement, the purchaser
42
must specifically advise the manufacturer that filters be placed at the input and output of
UPS systems. It is imperative that an investigation of the UPS equipment specification
sheet be made prior to purchasing the unit in order to make certain these filters can be
added by the manufacturer. Otherwise, satisfaction with the UPS will not be as desired
since these filters obviously do make the product more reliable.
This technical equipment specification is from Section 16611 (Facilities,
Electrical Components) from MASTERSPEC DRAWING COORDINATION, The
American Institute of Architects (AIA), December 2000. A user of UPS equipment can
modify this specification for their own uniquely personal situations. This particular
specification is directed toward a single-phase, on-line, static-type, Uninterruptible Power
Supply (UPS) system.
The data provided in the thesis shows conclusively that a small to mid-size UPS
can be modified to prevent the occurrence of harmful electromagnetic interference at
radio-receiving and sensitive data-processing facilities. Furthermore, this can be
achieved by the use of inexpensive commercial components.
43
THIS PAGE INTENTIONALLY LEFT BLANK
44
IV.
A.
SUMMARY AND RECOMMENDATIONS
SUMMARY
Electromagnetic interference (EMI) can degrade the operation and performance of
many kinds of electronic devices and radio receiving systems. This is especially the case
at Department of Defense radio-receiving sites and sensitive data-processing facilities.
The increasing use of computers, digital-data-processing devices, and power-control
devices in such facilities [including Uninterruptible Power Supplies (UPS), switching
power supplies, and motor controllers based on solid-state switching techniques] has
resulted in many cases of EMI problems. This thesis concentrated on EMI generated by
UPS.
EMI is generated by the rectifier and inverter sections of UPS systems, which
often produce noise up to and sometimes higher than 50 MHz. Also, another intermittent
source of impulsive noise is load-current changes (load switching) that create voltage and
current impulses and electrical noise.
The EMI generated by two unmodified commercial models of UPS was examined. The detailed results from one of the two systems are provided in this thesis. The
second system provided almost identical results. The particular units available for testing
were obtained for another task, and they were selected because of their low ambient
levels of EMI compared to other available commercial units. Nevertheless, their
conducted EMI levels were considered too high for use in HF and VHF radio-receiving
sites and for some sensitive data-processing uses.
The units were modified in accordance with integrated Barrier, Filter, Ground
techniques and again tested. Significant reductions in conducted EMI current was
achieved on all conductors penetrating the UPS case including all power and ground
conductors. The modifications resulted in the test units meeting all known conducted
45
EMI current limits including those of MIL-STD-461 and the suggested limits provided by
the US Navy SNEP team.
The MASTER DRAWING COORDINATION document issued by the American
Institute of Architects (AIA) dated December 2000 is often used as a guide for the
procurement and installation of UPS systems. This document, in its original form, does
not consider the impact of UPS-generated EMI on other systems. Based on the results of
this investigation, suggested changes to the AIA document are provided in Appendix F.
B.
RECOMMENDATIONS
The Naval Security Group (NSG) issued a recommendation in 1994 for the use of
UPS, which is shown below. This recommendation is fully supported, and the DOD will
be best served by following the NSG recommendation:
“ONLY MISSION-ESSENTIAL EQUIPMENT NOT ABLE TO TOLERATE
EVEN MOMENTARY POWER DISTURBANCES SHALL BE CONNECTED
TO UPS POWER.”
[From NSGINT 113`0.1D G43 (CRITICAL LOAD), 27 Jan 1994
In addition, to the above general recommendation, additional precautions need to
be taken to ensure that UPS-generated EMI does not degrade the operation of other radio
and electronic systems. This is especially the case in HF, VHF, and UHF radio-receiving
sites where UPS-generated EMI is often found at the input terminals of the radio
receivers and in sensitive data-processing facilities where UPS-generated EMI is often
found on data-cable shields and grounds.
When a determination is made that an UPS is required, or an UPS already exists,
specific guidelines and critical aspects for consideration both prior to purchase and after
the product has been obtained or installed should be followed. Following are
precautionary measures and information designed to insure product satisfaction and to
guide a purchaser through the ordering process:
46
♦ Upon determining the feasibility of purchasing an UPS, it must be
remembered that the UPS will use energy in a rather inefficient manner,
although at times (such as in a mission critical area) the benefit provided by an
UPS far outweighs any detrimental aspect.
♦ Although the electronics, mechanical components, and battery portions are
quite reliable, an UPS will require regular maintenance or repair. Because of
this, an annual maintenance contract included with any order is desirable to
insure reliable operation and optimum satisfaction.
♦ Prior to purchase, a determination should be made as to whether any location
chosen for installation is a sensitive facility (i.e., data- processing or radioreceiving facility), due to the fact that these areas have the highest
susceptibility to noise and interference.
♦ Care should be taken to review the equipment specification sheet, to ensure
that EMI filters have been properly installed by the manufacturer.
♦ In most cases, the use of EMI filters on input and output power conductors
will be an additional item to purchase; however, these filters are considered
"off the shelf" and in actuality it will prove less costly to have them installed
at the manufacturer's plant prior to shipment.
♦ Upon arrival of the order, and before acceptance of the UPS, carefully peruse
the accompanying equipment specification sheet. This provides an
opportunity for equipment modifications to be made in advance of acceptance.
Although an UPS provides excellent protection for sensitive equipment from
many kinds of power faults, the generation of low- and high-frequency EMI must be
considered. At times, interference may exist after the UPS has been installed. During
these occasions, the procedure for identifying and documenting harmful levels of EMI
becomes somewhat more complicated. The following recommendations or suggestions
should be considered after installation has been accomplished and EMI is an identified
problem:
♦ A study must be conducted by a team of EMI technical experts to investigate
all sources of harmful levels of EMI.
♦ The team must narrow the cause of EMI from many potential sources to the
UPS.
47
♦ Prior to taking any action to modify the UPS, all other sources of EMI must be
considered. The time of day/week when the EMI occurs at the site should be
identified to determine if a pattern of occurrence is repetitive, as this is a key
in identifying the sources of EMI.
♦ Additionally, improper grounding or radiated EMI may also be a reason for
interference problems.
♦ If an UPS is shown to be a source of EMI, it is recommended that wide-band
current clamps and a spectrum analyzer be used to examine the level of the
EMI over frequency ranges of interest.
♦ If EMI current exceeds the threshold of criteria provided by the SNEP teams
or the limits in MIL STD 461, then filters should be added to all input and
output conductors of the UPS system. Appendix E identifies several steps to
determine the filter specification decision making process.
In cases where UPS noise may be a factor degrading the performance of
electronic and radio devices and systems, it is recommended that UPS be procured and
installed in accordance with the additions to “The MASTER DRAWING COORDINATION document issued by the American Institute of Architects (AIA) dated December
2000”. These additions and changes are provided in Appendix F. These changes apply
to small and medium power UPS systems (up to about 50 kVA) where standard filters are
available as COTS items.
Finally, it is recommended that future projects be undertaken to better understand
methods to control EMI caused by a medium-sized (i.e., 50 to 250 kVA) UPS and large
sizes (i.e., 250 kVA and higher). The project is required to develop methods and filter
configurations to lower conducted and radiated EMI from medium and high power UPS
to acceptable limits.
48
APPENDIX A.
SOURCES OF EMI AND NOISE
This Appendix provides a “Table of Noise Sources” obtained from United States
Signals Intelligence Directive dated July 1998. The title of the directive was called
“Electromagnetic Compatibility Technical Guidelines”.
The sources in Table A.1 are intended only for general guidance to planners and
operational units since it is impossible to list all possible sources. While only a limited
number of sources existed in past years, the recent introduction of new devices into
Department of Defense facilities, especially digital power-control devices, has introduced
many new kinds of sources. Each situation involving electrical and radio-noise problems
must be evaluated on a case-by-case basis to identify and mitigate any adverse input on
facility operations from each individual source.
Existing noise sources in Department of Defense facilities cannot always be
summarily removed because of overall operational considerations, but each source can
and should be modified or replaced to minimize the deleterious impact of noise on the
operation of electrical and electronic systems. Every effort must be made to protect
against ongoing mission performance degradation created by EMI. Such actions will
improve the performance of radio and data-processing systems and usually result in the
improved efficiency of the devices causing noise.
Of special concern is the recent introduction of COTS equipment into DOD
facilities without full consideration of the possibility that some COTS devices are major
sources of radio and electrical noise. Once such devices are introduced into a facility, a
major effort is often required to reduce EMI from them to harmless levels.
Table A.1 includes many sources of EMI identified in past years. It includes
sources producing both radiated and conducted EMI.
49
Table A.1: NOISE DESCRIPTIONS AND TYPICAL SOURCES
SOURCES
NOISE DESCRIPTIONS
Diesel engine generators
Transformers
Battery charger
Rarely produce radio noise
Rarely produce radio noise
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
Mercury arc rectifiers
Silicon-controlled rectifiers
(SCRs)
Separate radio ground and
power
Ground in dry weather
Hot or burned fuse holders
Hot or burned switch or
circuit breaker contacts, wet
insulation, wires on trees,
etc.,in wet weather
Faulty heating devices
Incandescent lamps, with
broken filament or loose in
socket
Thermostats
Voltage regulators
Slowly opening switches or
Controllers
Certain motors during starting
Period
Faulty heating device
Fluorescent lamps
More or less continuous frying
noise hum; harmonics with power
More or less continuous frying
noise hum; harmonics with power
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
50
Table A.1 (Continued): NOISE DESCRIPTIONS AND TYPICAL SOURCES
Ultraviolet-ray machine
Diathermy machine
Small motors with commutator,
such as electric razors, drills,
vacuum cleaners, or mixers
Vehicle or variable speed
commutator motor
Stationary gasoline engine
Switches being turned off and
on
Appliances being plugged in or
disconnected
Thermostats
Time clocks
Telephone dialing
Ungrounded wires or pieces of
sheet metal blowing against
other pieces of metal in windy
weather
Electric fence
In dry weather - "high-line
noise”, usually off site
In dry weather - neon sign,
usually off site
During snow storm, or sand
storm, sometimes with light
rain, but always with windprecipitation static corona
effects from ungrounded guy
wires, antennas, or insulated
metal structures
Frying noise with power hum
harmonics; cut off and on
Frying noise with power hum
harmonics; cut off and on
Buzzing noise
Buzzing noise
Erratic clicks of different intensities
Erratic clicks of different intensities
Erratic clicks of different intensities
Erratic clicks of different intensities
Erratic clicks of different intensities
Erratic clicks of different intensities
Erratic clicks of different intensities
Regular clicks (frequently at
1-second intervals)
Faint hiss with or without power
frequency - hum modulation,
punctuated by a "sleet on tin roof"
effect
Faint hiss with or without power
frequency - hum modulation,
punctuated by a "sleet on tin roof"
effect
Faint hiss with or without power
frequency - hum modulation,
punctuated by a "sleet on tin roof"
effect
51
THIS PAGE INTENTIONALLY LEFT BLANK
52
APPENDIX B.
A.
EMI RELATED TECHNICAL STANDARDS
INTRODUCTION
Many organizations and agencies provide standards, handbooks, and other
publications related to EMI. Following is a partial listing of such documents:
♦ FCC Part 15, “Radio Frequency Devices”, Federal Communications
Commission, Washington D.C.
♦ FCC Part 18, “Industrial, Scientific, and Medical Equipment”, Federal
Communications Commission, Washington D.C.
♦ ANSI/IEEE, “Radio Interference: Methods of Measurement of Conducted
Interference Output to the Power Line from AM and Television Broadcast
Receivers in the Range of 300 kHz to 25 MHz”, American National
Standards Institute (ANSI) and Institute of Electrical and Electronic
Engineers (IEEE).
♦ ANSI/IEEE 214, “Construction Drawings of Line Impedance Networks
Required for Measurements of Conducted Interference to the Power Line
from FM and Television Broadcast Receivers in the Range of 300 kHz to
25 MHz as Specified in ANSI/IEEE Standard 213”, American National
Standards Institute (ANSI) and Institute of Electrical and Electronic
Engineers (IEEE).
♦ ANSI/IEEE 518, “Guide for the Installation for Electrical Equipment to
Minimize Noise Inputs to Controllers from External Sources”, American
National Standards Institute (ANSI) and Institute of Electrical and
Electronic Engineers (IEEE).
♦ ANSI/IEEE C63.2, “ANSI Specification for Electromagnetic Noise and
Field Instrumentation, 10 kHz to 40 GHz”, American National Standards
Institute (ANSI) and Institute of Electrical and Electronic Engineers
(IEEE).
♦ ANSI/IEEE C63.4, “ANSI Specification for Radio-Noise Emissions from
Low-Voltage Electrical and Electronic Equipment in the Range of 10 kHz
to 1 GHz”, American National Standards Institute (ANSI) and Institute of
Electrical and Electronic Engineers (IEEE).
♦ ANSI/IEEE C63.12, “Recommended Practice of Procedures for Control of
System Electromagnetic Compatibility”, American National Standards
Institute (ANSI) and Institute of Electrical and Electronic Engineers
(IEEE).
53
♦ MDS-201-0004, “Electromagnetic Compatibility Standard for Medical
Devices”, Federal Drug Agency (FDA) Regulations.
♦ NFPA 70, “National Electric Code”, National Fire Protection Association
(NFPA).
♦ NFPA 70E, “Electrical Safety Requirements for Employee Workplace",
National Fire Protection Association (NFPA).
♦ NESC Handbook, “National Electrical Safety Code Handbook”, Edition 4,
Institute of Electrical and Electronic Engineers (IEEE).
♦ “United States Signals Intelligence Directive", July 1982.
♦ NSG Instruction (NSGINST) 2450.1, NAVSECGRU “Shore Electronics
Criteria”, Naval Security Group, Fort Meade, MD.
♦ MIL-HDBK-1004, Military Handbook, “Preliminary Design
Considerations”, U.S. Dept. of Defense
♦ MIL-HDBK-419A, Military Handbook, “Grounding, Bonding and
Shielding for Electronic Equipment and Facilities”, U.S. Dept. of Defense,
December 1987.
♦ MIL-STD-461E, Military Standard, “Requirements for the Control of
Electromagnetic Interference Characteristics of Subsystems and
Equipment”, (Replaces previous editions of 461 and 462), U.S. Dept. of
Defense, August 1999.
♦ DOD C-3222.5, “Electromagnetic Compatibility (EMC) Program for
SIGINT Sites”, U.S. Dept. of Defense, July 1988.
♦ IEEE Standard 1100-1999, “IEEE Recommended Practice for Powering
and Grounding Electronic Equipment”, IEEE Emerald Book, 1999
♦ FIPS Publication 94-1983, “Guideline on Electric Power for ADP
Installations”, NTIS, U.S. Department of Commerce.
♦ CISPR 22, “Limits and Methods of Measurement of Electromagnetic
Disturbance Characteristics of Information Technology Equipment (ITE)”,
1985.
♦ IEC-61000-4-6, “Immunity to Conducted Disturbances”, International
Engineering Consortium, 1995.
The large number of handbooks, standards, and documents related to EMC is an
indication of the importance of the topic as well as the diverse nature EMC. No other
topic related to the use of electricity has generated such a large list of such documents as
well as a massive number of articles in the technical journals and other publications.
54
Many of the above publications are large and complex, and they require
considerable technical expertise and costly equipment to comply with their requirements.
In addition, it is not always clear which document applies to specific cases, how to handle
situations introduced by new technology and how to resolve conflicts between the various
publications.
Of special interest is the recent tendency to specify the purchase of electrical and
electronic equipment and devices based on Commercial-Off-The-Shelf (COTS) requirements. In many cases this has resulted in the ability to quickly obtain new technology
such as advanced computers and data-processing devices. In other cases, this process has
resulted in the introduction of severe EMI problems into government and commercial
facilities (for example, the use of variable-speed motor drives in and around radio
receiving sites that are purchased to COTS requirements).
B.
DISCUSSION
It is not feasible to provide a comprehensive review of each of the listed
standards, handbooks, and documents in this Appendix. This is a formidable task that
needs to be done to highlight the inconsistencies and even some errors that are buried in
the available documents. Many of the listed documents are recent editions based on older
versions that contain some information of little value and some that is quite misleading.
Some of the documents even contain information that is technically incorrect.
To obtain a partial understanding of these matters, partial reviews of selected
documents are provided. The reader must recognize that the reviews provided are very
limited in scope and they do not provide a good overall evaluation of the applicability of
the selected document to specific cases. Even though the author researched the above
documents, only the most pertinent documents are described below.
55
C.
DOCUMENTS
1.
American National Standards Institute (ANSI) 63.12
ANSI 63.12 is titled the "American National Standard Recommended Practice on
Procedures for Control of System Electromagnetic Compatibility". Figure B.1 shows the
conducted emission guidelines.
COMMON-MODE CONDUCTED EMISSION
GUIDELINES (ANSI C63.12)
120
Current (dBµ A)
100
80
60
40
20
0
0.01
0.1
1
10
100
Frequency (MHz)
Figure B.1: ANSI C63.12 Conducted Emission Guidelines.
The reader should note that the vertical axis (or current) is in terms of dBµΑ. The
decibel is used extensively in electromagnetic measurements. The “dB” is the logarithm
of the ratio of two amplitudes. Examples of amplitudes are power, voltage, current,
electric field units, and magnetic field units. The power ratio is:
decibel = dB = 10 log ( P2 P1 ) .
(B-1)
Measurements can be expressed in terms of current ratios. In this case, replace P with
I2R. If the impedances (50 ohms) are equal, the ratio for current becomes:
56
dB = 20 log ( I 2 I1 )
(B.2)
Since the EMI current measured is in terms of dBmA and dBµA, the equation
becomes:
dBmA = 20 log ( I 2 I1 ) − 30
(B-3)
dBµ A = 20 log ( I 2 I1 ) − 60.
(B-4)
Included in this section is the power-to-current equation associated with this
thesis. The power-to-current equation for 50 ohms is:
dBµA = dBm +107.
(B-5)
The acceptable guideline for the threshold of interference is to remain at
approximately 3 mA or 69.5 dBmA. The guideline is cut off at 30 MHz since conducted
emission is generally negligible above the 30 MHz, owing to line losses. The proposed
guideline as shown in Figure B.1 is described by Table B.1.
Table B.1: Common-Mode Conducted Emission Guidelines
Frequency of Emissio
2.
Common-Mode Current
Below 800 kHz
2400 / f (kHz) mA
Above 800 kHz
3 mA
FCC Regulations: Class B Conducted Limits
The FCC conducted limits for FCC Class B non-intentional emitters are specified
in FCC Section 15.107 "Conducted Limits" Paragraph a. The limit for Class A devices
is: “For equipment that is designed to be connected to the public utility (AC) power line,
the radio frequency voltage that is conducted onto the AC power line on any frequency or
frequencies within the band 450 kHz to 30 MHz shall not exceed 250 microvolts.”
57
Assuming measurement across a 50-ohm load, this is equivalent to 5 µA. This
was based on 250 µV/50 Ω = 5 µΑ. This is equivalent to 14 dBµA (20 log (I)), where I
is in terms of µΑ.
3.
MIL-STD-461E (CE-102) Conducted Emissions
MIL-STD-461E (CE-102) conducted emissions specification specifies 60 dBµV
or 20 µA (or 26 dBµA) for "28 VDC" power supply leads. The limit is reduced by 6 dB
for 115 VAC power lines to 32 dBµA. However, MIL-STD-461E (CE-102) only
specifies the current limit to 10 MHz.
MIL-STD-461E also gives the emission and susceptibility requirements for
conducted type EMI. Figure B.2 shows its conducted current limits from 60 Hz to 100
kHz.
M IL -S T D -461E (C S 109 L im it)
140
Current (dBµA)
120
100
80
60
40
20
0
10
100
1000
10000
100000
Frequency (Hz )
Figure B.2: MIL-STD-461E, CS109 Limit
4.
US NAVY SNEP Documentation
An additional published source has been documented by the US Navy Signal-toNoise Enhancement Program (SNEP). Suggested limits have been established for EMI
58
current injected into ground conductors and all other related conductors of a receiving
site. The SNEP teams have recommended various EMI standards for equipment installed
in data-processing and signal-receiving sites [Reference 6]. These limits are provided in
Tables B.2 and B.3.
Table B.2: Suggested Maximum Permissible Limits for
Conducted EMI Current for Large Receiving Site.
Frequency Range
Maximum Current
0 to 10 kHz
2 mA
100 kHz to 100 MHz
10 µA
Table B.3: Suggested Maximum Permissible Limits for
Conducted EMI Current for Small Receiving Site.
Frequency Range
Maximum Current
0 to 10 kHz
2 mA
100 kHz to 100 MHz
2 µA
These limits have been established from extensive field measurements of the
susceptibility of HF, VHF, and UHF receiving sites to internal sources of EMI by US
Navy Signal-to-Noise Enhancement teams. The maximum limit between 10 kHz and 100
kHz is established by linear extrapolation from the disparate limits. When these limits
were met at radio receiving facilities, no interference from internal sources of EMI was
encountered.
While suggested maximum levels of EMI current have been provided for
receiving sites, similar levels for data-processing sites have not yet been established.
59
THIS PAGE INTENTIONALLY LEFT BLANK
60
APPENDIX C.
A.
DESCRIPTION OF UPS SYSTEMS
GENERAL
An Uninterruptible Power Supply (UPS) is used to provide electric power to
critical loads when utility power is interrupted, removed, or fails. Figure C.1 shows a
diagram of a typical single-module UPS system. A battery is trickle charged from the
AC-to-DC converter, usually with a full-wave rectifier. The battery drives an inverter
using transistors for lower-power units and thyristors or silicone-controlled rectifiers
(SCRs) as solid-state switchers for higher-power units. The details and configurations of
UPS systems, their performance and capacity, and their physical configuration vary
somewhat from one manufacturer to another, but all provide electrical power to loads for
a short period of time when their input power fails. UPS units also provide protection
against a number of other power problems including surges, sags, transients, dropouts
and brownouts as well as total power failures. The duration of this protection is
dependent on a combination of battery capacity and load. For times exceeding the
protection provided by an UPS, and where operation is critical, standby diesel enginedriven generators must be provided.
UPS MODULE
AC POWER
SOURCE
L
L
N
DC/AC
AC/DC
SENSITIVE
EQUIPMENT
N
RECTIFIER/
CHARGER
INVERTER
G
G
BATTERY PACK
Figure C.1:
Typical On-Line Single-Phase Single-Module UPS Diagram.
61
B.
UPS COMPONENTS
An UPS is a device used to provide continuous, acceptable power to its load
regardless of the input power supplied. UPS systems come in various types and sizes.
An UPS is used to provide clean, conditioned, and continuous power to critical electronic
equipment. It can protect electronic equipment from most detrimental conditions
experienced on power systems. The UPS will provide power to the load including an
event causing a total outage. The UPS supplied power will last only as long as the
systems battery bank will permit, typically 10 minutes. These systems are commonly
used to protect computer systems from short-term outages such as those experienced
during stormy conditions or other inclement weather situations.
Figure C.1 depicts the major components of a typical single-phase single-module
UPS system. The UPS systems tested for this thesis are of a type shown in Figure C.1.
They had a single input with a maintenance bypass section. The components of this UPS
system contained the following sections:
♦ Rectifier or Charging Unit - takes the utility AC power and converts it to DC
and also charges the batteries.
♦ Inverter - takes the DC from the rectifier or batteries and converts it to AC for
use in the computer system.
♦ Battery Bank - supplies DC power for the inverter in the event of
unacceptable AC input.
C.
UPS STANDARD
While no specific government or industry standard exists for the specification of
an UPS system, two sources provide useful information to aid in their design,
procurement and use.
Underwriters Laboratories (UL) Inc. publishes Standards for Safety documents.
Their document UL 1778 titled "Uninterruptible Power Supply Equipment", provides
recommendations about construction, performance, rating, marking, and testing of UPS
systems. The following sections in UL 1776 have applicable references to this thesis:
62
♦ Section 42: The leakage current that is accessible to the user shall not be more
than 0.75 milliampere.
♦ Section 42.1a: Leakage current shall not exceed 5.0 milliamperes.
♦ Section 48.2: Testing criteria on UPS systems that have EMI filter capacitors
installed.
♦ Section 74.1.2.j: For an UPS to have circuit filtering to meet EMC/EMI
regulations.
The Federal Communications Commission (FCC) specifies EMI limitations for
electronic and electrical devices under Part 15 of their regulations. They provide
maximum permissible levels of EMI voltage on the conductors supplying power to
electronic devices intended for use in industrial and residential applications. The
maximum permissible levels for industrial (Class A level) use are much higher than for
residential (Class B level) use. The FCC specified limitations are often used for the
procurement of COTS UPS systems by the government. No mention is made in the FCC
documentation of the use of UPS systems at radio-receiving or data-processing facilities.
D.
UPS INVERTER SECTION
Most inverters in UPS products employ solid-state switching techniques to
convert the DC power to AC power. Some of the techniques employed in the conversion
are square-wave, stepped-wave or pulse-width modulation. Each technique has
beneficial qualities, but each has limitations. The detrimental aspects of the switching
actions of an inverter are output distortion of the voltage waveform, limitations in
transient response, efficiency, and the generation of EMI. This aspect of an UPS
converter should be taken into account when selecting a unit for installation in a facility.
63
E.
UPS HAZARDOUS WARNING LABELS
Figures C.2 and C.3 show EMI/RFI warning labels found on the external surface
of the case of two UPS systems installed at two different receiving sites. These labels are
also reproduced in the manuals of both systems. Such warnings are required in the
United States on all commercial electronic and electrical equipment that generates
EMI/RFI. Both of the UPS systems containing the warnings were purchased as COTS
devices in accordance with FCC Class A requirements. This is a common procurement
procedure since UPS systems meeting the stricter Class B requirements are seldom
specified during procurement, and they are seldom available from manufacturers as
COTS devices.
WARNING: This equipment generates, uses, and can radiate radio frequency
and if not installed and used in accordance with instructions may cause interference
to radio communications. It has been tested and found to comply with the limits for
Class A computing devices pursuant to Subpart J or Part 15 of FCC Rules, which
are designed to provide reasonable protection against such interference when
operated in a commercial environment. Operation of this equipment in a residential
area is likely to cause interference in which case the user at his own expense will be
required to take whatever measures may be necessary to correct such interference.
CAUTION:
Always be aware that hazardous voltages may be present within the UPS even
when the system is not operating.
Figure C.2:
LABEL ON AN UPS INSTALLED IN A RECEIVING SITE1
This equipment complies with the requirements in Part 15 of FCC Rules for a Class
A computing device. Operation of this equipment in a residential area may cause
unacceptable interference to radio and TV reception requiring the operator to take
whatever steps are necessary to correct the interference.
Figure C.3:
Another UPS Warning Label2
1
This label was reproduced from a photograph taken from a UPS located at NSGA Northwest during a
visit to the site in April 1997. The photograph is located in field notebooks for that visit.
2
This label was reproduced from a photograph taken from an UPS located at Detachment L, Field Site
Korea during a visit to that site in October 1999. The photograph is located in field notebooks for that visit.
64
The warning labels clearly indicate that EMI/RFI can be expected if a Class A
UPS is located at a facility containing radio receivers. This should be sufficient warning;
the additional steps will probably be required to ensure EMI/RFI problems are not
encountered after a Class A UPS is installed and becomes operational. Unfortunately,
very costly corrective actions to the UPS must be implemented to reduce EMI/RFI to
harmless levels after the installation.
No mention is made in the warning labels of possible adverse affects of EMI
generated by a Class A UPS system to data-processing devices, the possible
contamination of signals carried by wired local area networks (LANs) or other adverse
effects to electrical or electronic devices.
The procurement and installation of Class A UPS systems require attention during
the design stages of a facility and prior to procurement. This is especially the case for
UPS systems intended for use in or near facilities containing HF, VHF and UHF radio
receivers. Class A devices of any kind should not be used at or near receiving facilities
without first ensuring that problems will not be encountered or all Class A devices are
modified to reduce RFI/EMI to harmless levels.
65
THIS PAGE INTENTIONALLY LEFT BLANK
66
APPENDIX D.
A.
CALIBRATION
GENERAL
Normal performance tests and system calibrations were made at the start of each
set of measurements or at any time a significant change was made in the instrumentation.
Of special concern was the transfer of amplitude calibration scales from the HP 141
Spectrum Analyzer to the ELF Model 7200B 3-Axis Display. In addition each current
probe was calibrated with test fixtures specifically designed for that purpose.
B.
TIME-AXIS CALIBRATION OF THE 3-AXIS DISPLAY
The duration of the time axis of the time-history views is dependent on the scan
time of the spectrum analyzer and the blanking or dead time at the end of each scan. The
time axis can be calculated from the following formula:
Tt = (ts + tb ) ⋅ 60
(D-1)
where:
Tt is the duration of the time axis in seconds,
ts is the total time of each scan of the spectrum analyzer in seconds, and
tb is the blanking or dead time between each scan in seconds.
An initial value of ts can be obtained from the scan-time control on the spectrum
analyzer. Additionally, this can be measured with an oscilloscope or a counter to obtain
more accurate values. The value of tb is not provided by the manufacturer, and it must be
measured with an oscilloscope or a counter. The value of tb varies from analyzer to
analyzer and with the setting of the scan-time control.
Table D-1 shows a typical calibration chart summarizing the parameters needed to
establish the time-axis values when the scan-time of the spectrum analyzer is operated in
the auto-sync mode.
67
Table D-2 shows a typical calibration chart summarizing the parameters needed to
establish the time-axis values when the scan-time of the spectrum analyzer is operated in
the line-sync mode.
Table D-1: Time Scale Calibration, Auto Sync
Scan Time
(in ms)
2
5
10
20
50
100
200
500
1000
2000
5000
Measured Scan Time
(in ms)
1.9
4.3
9.75
20
45
105
200
450
1025
2050
4600
Blanking Time
(in ms)
0.29
0.30
5.0
5.0
6.0
85
85
85
85
675
688
Time Axis
(in sec)
0.13
0.28
0.89
1.5
3.1
11.4
17.1
32.1
66.7
163
317
Table D-2: Time Scale Calibration, Line Sync
Scan Time
(in ms)
2
5
10
20
50
100
200
500
1000
2000
5000
Measured Scan Time
(in ms)
2.45
4.5
10
20
45
105
200
450
1025
2050
4600
68
Blanking Time
(in ms)
Time Axis
(in sec)
15
1.0
12.5
6.8
13
22
95
100
90
95
937
687
1.0
1.0
2.0
4.0
12
18
32.4
67
179
317
C.
AMPLITUDE CALIBRATION
The amplitude calibration of the spectrum analyzer must be transferred to the
time-history display. The amplitude calibration levels of the spectrum analyzer are
recorded on the time-history display in 10-dB steps starting with the full trace setting on
the analyzer and proceeding downward in steps. A photograph is then made of the
amplitude traces on the time-history display. It is important that the same camera used to
photograph the calibration traces is used for all subsequent data-recording work. This
avoids camera-to-camera variations.
Figure D.1 shows an example of an amplitude-calibration photograph. The levels
shown on the original photograph can be transferred to any example of measured data to
obtain amplitude scales in mA or µA as needed. Since the dynamic range of the
measured data was very large, it was necessary to use a logarithmic amplitude scale to
portray the full range of EMI current.
Figure D.1 Amplitude Calibration
69
D.
BANDWIDTH CALIBRATION
EMI produced from the solid-state switches in an electronic UPS is highly
impulsive, resulting in spectral components of EMI that are much wider than the
bandwidth of the measurement process available from a spectrum analyzer. Because of
this, the amplitude of impulsive noise is a function of the measurement bandwidth. A
calibration curve permitting the conversion of impulse amplitude from the measured
value to other bandwidth values is provided to aid in the analysis and application of the
data. Figure D.2 provides the curve used to aid in the evaluation of the data obtained and
shown in this document. A reference curve to scale changes in the amplitude of timestable gaussian noise with bandwidth is also provided.
Figure D.2
E.
Bandwidth Scaling Curve
PROBE CALIBRATION
The two current probes used to measure EMI current levels were flat in frequency
response over their normal operating frequency range. A Tektronix Model 6021 Current
Probe interfaced with a Tektronix Model CT-4 Current Probe was used to measure EMI
70
current at frequencies below about 100 kHz. While the specified frequency range of this
combination of probes extended to much higher frequencies (up to 50 MHz), the higher
portion of its range could not be used. The limited dynamic range of the instrumentation
(about 80 dB) prevented the measurement of low-level spectral components of current
above a few tens of kHz. Nevertheless, the combination of the P6021/CT-4 probes
provided a means to understand harmonic content of the input and output power of UPS
systems and the spectral components of low-frequency EMI current. It was necessary to
provide a means to convert the probe readings into amplitude scales in mA or µA on the
recorded data. The next charts (Figures D.3 and D.4) are calibration curves associated
with the current transformer (CT) used for the testing. The CT-4/P6021 current
transformer was used for the low-frequency test setup and F-70 CT was used in a highfrequency test setup.
The response of the P6021/CT-4 combination of probes falls at frequencies below
200 Hz, and a calibration curve is required to obtain compensation values of current
below this value. Figure D.3 shows the calibration curve used to understand lowfrequency current levels.
Figure D.3
CT-4/P6021 Probe Calibration
71
A Fischer Model F-70 Current probe was used to measure EMI current at
frequencies above 100 kHz. Its low-frequency response diminishes below 100 kHz and
allows the measurement of high EMI current levels at higher frequencies without concern
for the high levels of low-frequency current flowing on conductors. The response of the
probe was flat from 100 kHz up to 100 MHz. This permitted the use of a single
amplitude scale on the data describing high-frequency EMI current levels.
A calibration curve was used to extend the amplitude response of the F-70 probe
to frequencies below 100 kHz. This provided a means to extend the frequency range of
the F-70 probe downward into the useful frequency range of the CT-4/P6021 probe for
comparative measurements. Figure D.4 shows the calibration curve for the probe used
during these measurements.
Figure D.4
F-70 Probe Calibration Curve
72
APPENDIX E. POWER LINE FILTERS
A.
INTRODUCTION
The information in this section is limited to a description of filters composed of
discrete components; usually capacitors, inductors, and resistors. This is the type
normally used to correct EMI problems associated with the power conductors of
electronic equipment. Four main types of such filters can be obtained which are lowpass, high-pass, band-pass, and band-trap. For the purposes of this thesis, the discussion
will be limited to the low-pass type which allows alternating electrical power to be
provided to an uninterruptible power supply (UPS) while attenuating higher-frequency
EMI being conducted out of the UPS. Similarly, a second filter can be provided on the
output conductors to attenuate high-frequency EMI while allowing low-frequency
electrical power to be applied to a load. When properly installed at the surface of the
conducting case of an UPS, low-pass filters can be highly effective in reducing conducted
EMI current on entry and exit power conductors to harmless levels.
A number of references provide detailed information about the detailed design of
filters, and this readily available information is not duplicated in this appendix.
Fortunately, a variety of low-pass filters are available as standard catalog items up to
modest power-handling ratings. Two types were considered for use in the tests of the
UPS described in this thesis. The first was the standard “Pi” configuration shown in
Figure E.1.
73
Figure E.1: “Pi” Configuration Filter
The “Pi” configuration provides low attenuation to all frequencies below its cutoff frequency and a high transfer impedance to all spectral components above the cutoff
frequency. In addition, it provides a low input and output impedance to all spectral
components above the cutoff frequency.
The second type considered was the standard “T” configuration shown in Figure
E.2. This type also provides low loss to spectral components below its cutoff frequency,
and it also provides high transfer impedance to all spectral components above its cutoff
frequency. It differs from the “Pi” configuration in that it provides high input and output
impedance to all spectral components above the cutoff frequency.
Figure E.2: “T” Configuration Filter
74
Transfer impedance, Zt, is defined as:
Zt =
Ein
I out
(E-1)
where:
Ein is the voltage of a spectral component of EMI produced by the UPS, and
Iout is the current of the selected component of EMI on the outside conductors.
The reason both configurations of filters were considered is that some switching
devices (especially motor controllers using similar switching techniques) used for the
conversion of electric power are sensitive to the above-band impedance of filters on the
input and/or output conductors. The type of UPS chosen for the work described in this
thesis was insensitive to such impedance problems which allowed either type of low-pass
filter to be used. The standard “Pi” configuration was chosen simply because it was
readily available at low cost.
One additional consideration was used in the selection and use of filters. This is
the physical configuration of the green-wire ground connection to the input and output
sides of a filter. It is necessary to select a filter case configuration that provides a
conducting path for EMI current flowing on the green-wire ground conductor to return to
its source within the UPS case. This was accomplished by the selection of a standard and
inexpensive COTS filter using a metal case that complied with the electrical
configuration shown in Figure 3.19 of the main body of the thesis. This configuration
provides low transfer impedance to low frequencies to meet the safety requirement of the
NEC while providing high transfer impedance to high frequencies. The high transfer
impedance at high frequencies is obtained from the shielding of the metal case housing
the UPS.
75
Two additional terms are used to evaluate filter effectiveness. They are commonmode (CM) and differential mode (DM) EMI. CM EMI flows on both the white and
black conductors of a 120-V power source, and CM EMI voltage is measured from either
conductor to ground. DM EMI flows in one direction on the white wire and in the other
direction on the black wire. DM voltage is measured between the black and white
conductors. Both CM and DM modes are considered in this thesis.
B.
FILTER CHARACTERISTICS
1. General
In order to select an appropriate EMI power-line filter, the following are some of
the characteristics that a user must be concerned with: type, application, performance,
agency approvals, insertion loss range, current ratings, temperature range and voltage
range. The EMI power-line filter is of a low-pass type that is used on electronic
equipment. The filter's function is to block the flow of EMI current while passing a
desired 60-Hz current.
2. Insertion Loss
Insertion loss (IL) is a measure for effectiveness of a filter. It is defined as the
ratio of voltage (E1) across the phase-to-ground, with a filter in the circuit at a given
frequency, while voltage (E2) across the phase-to-ground contains no filter in the circuit
at the same frequency. Since insertion loss is dependent on the source and load
impedance in which a filter is to be used, IL measurements are defined for a matched 50ohm system. The IL is measured in decibels (dB) and defined as:
E 
IL ( dB ) = 20 log  1  .
 E2 
76
(E-2)
The insertion filter loss equation is also expressed in terms of number of poles or
stages of filter:
  f 2 N 
IL ( dB ) = 10 log 1 +  1  
  f 2  
(E-2)
where f1 = interfering EMI frequency,
f2 = cutoff frequency of filter, and
N = number of poles or stages of filter.
The number of stages determines the attenuation slope of the filter's characteristic
curve. If the filter has 2 stages, then the slope of the curve is 40 dB per decade.
C.
FILTER STANDARDS
The following military and commercial standards documents associated with
filters and insertion loss measurements are:
•
MIL-F-15733: Filters and Capacitors, Radio Frequency Interference
•
MIL-F-28861: Filters and Capacitors, Radio Frequency Interference /
Electromagnetic Interference Suppression
•
MIL-STD-220: Method of Insertion Loss Measurement
•
ANSI C63.13: American National Standard Guide on the Application and
Evaluation of EMI Power-Line Filters for Commercial Use
When specifying in accordance with MIL-STD-220 the minimum insertion loss
shall be accordance with Table E.1.
Table E.1: MIL-STD-220 Insertion Loss Filter Characteristic
Frequency
AC Insertion
Loss (dB)
15 kHz
10
100 kHz
42
500 kHz
70
77
1 MHz
70
100 MHz
70
1 GHz
70
D.
INPUT FILTER
A Delta power-line EMI filter was used for purposes of this research, as the filter
is a common "off-the-shelf" item and available from the SNEP laboratory. The
schematic and characteristics are shown in Figure E.3. The insertion loss diagram (for a
DELTA ELECTRONICS, 10DRDG3, EMI FILTER) was obtained from the
manufacturer and is shown in Figure E.4.
E.
OUTPUT FILTER
The output filter used for thesis research was a Corcom EMI filter Model 6VSK7.
It is a 6-amp, 120-volt, 60-Hz, single stage filter.
F.
INSTALLATION CRITERIA
The proper installation of a filter is critical to achieve successful filtering of EMI
generated by an UPS or preventing EMI from external sources from affecting the
operation of an UPS. The filter must provide an electrical barrier to prevent harmful
spectral components of EMI from escaping or entering the UPS. To accomplish this, the
filter must be installed on the metal case of an UPS and in accordance with the principles
shown in Figure 3.19 of the main body of this thesis. Installation at other locations (i.e.,
in the interior of an UPS or externally on the power conductors of an UPS) will result in
the significant loss of effectiveness and the use of filters at such locations is inadvisable
and not recommended.
78
Manufacturer: Delta Electronics, Inc.
Model Number: 10DRCG3
DR Series, High Performance Filter
Two Stage Filter
Characteristics: 115v, 60 Hz, 40 degrees C, 10 amps
Filter is used in suppressing both line-to-line and line-to-ground noise.
All parts are UL recognized, CSA certified, and VDE approved
Specifications:
1) Maximum leakage current (line-to-ground) = 0.25 mA
2) Hipot rating (one minute):
line-to-ground = 2250 VDC
line-to-line = 1450 VDC
3) Operating frequency = 50/60 Hz
4) Rated voltage = 115/250 VAC
Where:
R= 2.2 MΩ
C1= .22 µF
C2= .22 µF
C3= .22 µF
Cy=3300 pF
L1, L2= 1 mH
Figure E.3: The Schematic and Characteristics of the Delta 10DRCG3 Filter
(Technical Product Catalog, Delta High Performance Filters, Delta Inc., Page 6-8,
Reference 7)
79
Figure E.4: Insertion Loss Diagram for the Input UPS Filter
(Technical Product Catalog, Delta High Performance Filters, Delta Inc., Page 6-8,
Reference 7)
When selecting an EMI power-line filter, several steps are recommended. They
are:
1. If a device is suspected of generating harmful levels of EMI, measure the EMI
current on all conductors entering and exiting the device. Ascertain if any
spectral component of current generated by the device appears to be abnormal.
2. Compare the measured levels of EMI current to emission limit criteria
provided by government publications or other organizations.
3. Determine the attenuation required to reduce EMI current levels to acceptable
limits.
4. Select a filter that provides sufficient attenuation and install it in accordance
with the principles provided in this thesis.
80
G.
FILTER MANUFACTURERS
Below in Table E.2 are several filter manufacturers from EEM 2001 (Reference
8). The EEM is well known electronic equipment source of technical information.
Information is also available from the EEM 2001 (http://eemonline.com) on the World
Wide Web.
Table E.2: Power Line Filter Manufacturers
AEROVOX
CORCOM, INC.
CURTIS INDUSTRIES
DEARBORN ELECTRONICS, INC.
DELTA ELECTRONICS, INC.
EMISSION CONTROL LTD.
FILTERS CONCEPTS, INC.
LINDGREN RF ENCLOSURES, INC.
MECHATRONICS, INC.
METUCHEN CAPACITORS, INC.
OKAYA ELECTRIC AMERICA
POWER DYNAMICS, INC.
RFI (DEL ELECTRONICS)
SAE POWER, INC.
SCHAFFNER EMC
SCHURTER, INC.
SPECTRUM CONTROL INC.
TAMURA CORP.
TEXAS SPRECTRUM ELECTRONICS, INC.
TRI-MAG INC.
WICKMANN USA
81
THIS PAGE INTENTIONALLY LEFT BLANK
82
APPENDIX F.
PROPOSED SPECIFICATION
Section 16611 from the MASTERSPEC DRAWING COORDINATION
document issued by the American Institute of Architects (AIA) and dated December 2000
provides an overall specification for an Uninterruptible Power Supply (UPS). The
specification is directed toward the purchase and installation of a single-phase, on-line,
static-type UPS.
This particular specification does not consider the possible impact of EMI
generated by an UPS on other electronic systems and devices, but users of UPS systems
in such facilities can modify this specification to meet their unique and special
requirements. Suggested changes in this specification are listed in bold. These changes
are directed at the use of UPS systems in radio-receiving and sensitive data-processing
facilities. Of particular concern is the elimination of harmful levels of UPS generated
EMI on radio signal reception and on the operation of sensitive data-processing systems.
______________________________________________________________
SECTION 16611 - UNINTERRUPTIBLE POWER SUPPLY
PART 1 - GENERAL
1.1
RELATED DOCUMENTS
A. Drawings and general provisions of the Contract, including General and
Supplementary Conditions and Division 1 Specification Sections, apply to this
Section.
1.2
SUMMARY
A. This Section includes 1-phase, on-line, static-type, uninterruptible power supply
(UPS) systems, complete with battery and battery circuit breaker.
83
1.3 DEFINITIONS
A. UPS: Uninterruptible power supplies that automatically provide power, without
delay or transients, during any period when normal power supply is incapable of
performing acceptably.
B. THD: Total harmonic distortion.
C. EMI: Electromagnetic Interference. EMI is the impairment of a desired
electromagnetic signal by an electromagnetic disturbance, i.e., electrical
noise. This undesirable electromagnetic emission or any electrical
disturbance, man-made or natural, which causes any undesirable response,
malfunctioning or degradation in the performance of electrical equipment
D. EMI filter: An EMI filter is a passive electronic device used to suppress
conducted interference present on any power or signal line. It may be used
to suppress interferences generated by the device. Most EMI filters include
electronic components to suppress both common- and differential-mode
interference.
1.4 SUBMITTALS
A. Product Data: Include data on features, components, ratings, and performance for
each product specified in this Section.
B. Shop Drawings: Detail fabrication, internal and interconnecting wiring, and
installation of UPS system. Include dimensioned plan, elevation views, and
details of control panels. Show access and clearance requirements. Differentiate
between field-installed and factory-installed wiring and components.
C. Product Certificates: Signed by manufacturers of UPS systems certifying that the
products furnished comply with requirements.
D. Qualification Data: For firms and persons specified in the "Quality Assurance"
Article.
E. Factory Test Reports: Comply with specified requirements. Include conducted
noise data over the frequency range of 1 kHz to 50 MHz.
F. Field Test Reports: For tests specified in Part 3.
G. Maintenance Data: For system and products to include in the maintenance
manuals specified in Division 1. Include the following:
84
1. Lists of spare parts and replacement components recommended to be stored at
the project site for ready access.
2. Detailed operating instructions covering operation under both normal and
abnormal conditions.
H. Warranties: Special warranties specified in this Section.
1.5 QUALITY ASSURANCE
A. Testing Agency Qualifications: An independent testing agency with the
experience and capability to conduct the testing indicated without delaying the
work, as documented according to OSHA criteria for accreditation of testing
laboratories, Title 29, Part 1907; or a full member company of the International
Electrical Testing Association.
1. Testing Agency's Field Supervisor: Person currently certified by the
International Electrical Testing Association or the National Institute for
Certification in Engineering Technologies, to supervise on-site testing
specified in Part 3.
B. Comply with NFPA 70.
C. Source Limitations: Obtain UPS, including components, from a single
manufacturer with responsibility for entire system.
D. Listing and Labeling: Provide UPS specified in this Section that are listed and
labeled as a factory-assembled unit.
E. Listing and Labeling: Provide UPS specified in this Section that are listed and
labeled for use in computer rooms. Comply with NFPA 75.
1. The Terms "Listed" and "Labeled": As defined in the National Electrical
Code, Article 100.
2. Listing and Labeling Agency Qualifications: A "Nationally Recognized
Testing Laboratory" as defined in OSHA Regulation 1910.7.
F. Comply with UL 1778.
85
1.6 DELIVERY, STORAGE, AND HANDLING
A. Deliver equipment in fully enclosed vehicles after specified environmental
conditions have been permanently established in spaces where equipment is to be
placed.
B. Store equipment in spaces with environments that are controlled within
manufacturer's ambient temperature and humidity tolerances for nonoperating
equipment.
1.7 WARRANTY
A. General Warranty: The special warranty specified in this Article shall not deprive
the Government of other rights the Government may have under other provisions
of the Contract Documents and shall be in addition to, and run concurrent with,
other warranties made by the Contractor under requirements of the Contract
Documents.
B. Special Warranty for Batteries: A written warranty, signed by manufacturer and
principal Installer, agreeing to replace UPS system storage batteries that fail in
materials or workmanship within the specified warranty period.
1. Special Warranty Period for Batteries: 10 years from date of Substantial
Completion. A full warranty applies to the first year of the period, and a
prorated warranty applies to the last 9 years.
1.8 EXTRA MATERIALS
A. Furnish extra materials described below that match products installed, are
packaged with protective covering for storage, and are identified with labels
describing contents. Deliver extra materials to Government.
1. Fuses: 1 for every 10 of each type and rating, but not less than 1 of each.
2. Cabinet Ventilation Filters: One complete set.
PART 2
PRODUCTS
2.1 MANUFACTURERS
A. Available Manufacturers: Subject to compliance with requirements,
manufacturers offering products that may be incorporated into the Work include,
but are not limited to, the following:
86
1. Best Power Technology, Inc.
2. Computer Power, Inc.
3. Controlled Power Co.
4. Deltec Corp.
5. Exide Electronics.
6. HDR Power Systems, Inc.
7. International Computer Power.
8. International Power Machines Corp.
9. Liebert Corp.
10. Mitsubishi Electronics America, Inc.
11. Pacific Power Source Corp.
12. Square D Co.; EPE Technologies, Inc. Subsidiary.
13. Toshiba International Corp.
2.2 MANUFACTURED UNITS
A. Description: Electronic components and switching devices are housed in one or
more metal cabinets, with batteries rack mounted separately. Automatic system
operating functions include the following:
1.
2.
3.
4.
5.
6.
7.
Normal Conditions: Supply the load with power flowing from the normal
AC power input terminals, through the rectifier/battery charger and inverter,
with the battery connected in parallel with the rectifier output.
Abnormal Supply Conditions: When the normal AC supply deviates from
specified voltage, waveform, or frequency limits, the battery supplies energy
to maintain constant inverter output to the load.
When normal power fails, energy supplied by the battery through the inverter
continues supply to the load without switching or disturbance.
When power is restored at the normal supply terminals of the system, the
rectifier/battery charger supplies power to the load through the inverter and
simultaneously recharges the battery. Synchronize the inverter with the
external source before transferring the load.
When the battery becomes discharged and normal supply is available, charge
the battery by the rectifier/battery charger. On reaching full charge, shift the
rectifier/battery charger to a float-charge mode.
When any element of the UPS system fails and power is available at the
normal supply terminals of the system, the static bypass transfer switch
switches the load to the normal source with less than one-quarter-cycle
interruption of supply.
If a fault occurs in the system supplied by the UPS and current flows in
excess of the overload rating of the UPS system, the static bypass transfer
switch operates to bypass the fault current to the normal supply circuit of the
UPS system for fault clearing.
87
8.
When the fault has cleared, the static bypass transfer switch returns the load
to the UPS system.
B. Functional Description of Manual Operation: Manual operating functions include
the following:
1. Turning the inverter off causes the load to be transferred by the static bypass
transfer switch directly to the normal AC input source without interruption.
2. Turning the inverter on causes the static bypass transfer switch to transfer the
load to the inverter.
C. Maintenance Bypass/Isolation Switch: Interlocked so UPS cannot be operated
unless the static bypass transfer switch is in the bypass mode. The device has 3
settings that produce the following conditions without interrupting supply to the
load during switching:
1. Full Isolation: Load is supplied bypassing UPS. UPS AC supply input, static
bypass transfer switch, and UPS load terminals are completely disconnected
from external circuits.
2. Maintenance Bypass: Load is supplied bypassing UPS. UPS AC supply
terminals are energized to permit operational checking, but system load
terminals are isolated from the load.
3. Normal: UPS AC supply terminals are energized and the load is being
supplied through either the static bypass transfer switch or the UPS rectifier
and inverter.
2.3 SYSTEM SERVICE CONDITIONS
A. Environmental Conditions: Operate continuously in the following environmental
conditions without mechanical or electrical damage or degradation of operating
capability:
1. Ambient Temperature: 0 to 40 deg C.
2. Relative Humidity: 0 to 95 percent, noncondensing.
3. Altitude: Sea level to 4000 feet (1220 m).
88
2.4
SYSTEM CHARACTERISTICS
A. Minimum Duration of Supply: 15 minutes, if rated full load is being supplied
solely from the battery.
B. System Performance When Supplied from Battery: Performance under steadystate and transient-load conditions remains within specified tolerances throughout
minimum duration of supply from battery specified.
C. Input Voltage and Frequency Tolerance: System steady-state and transient output
performance remains within specified tolerances when steady-state AC input
voltage varies plus or minus 10 percent from nominal voltage; when steady-state
input frequency varies plus or minus 5 percent from nominal voltage; and when
the THD of input voltage is 15 percent, and the largest single harmonic
component is a minimum of 5 percent of the fundamental value.
2.5
COMPATIBILITY WITH LOAD
A. Operate within specified performance tolerances, supply type of distribution system
indicated, and serve rated load comprised of various load elements. Load elements
provide an overall load profile with the following characteristics:
1. Aggregate Load for Single-Phase Electronic Equipment with Switch-Mode
Power Supplies, served at 120 or 208 V: 55 percent of UPS capacity.
2. Aggregate Load for Polyphase Electronic Equipment with Switch-Mode
Power Supplies, served at 208 V: 5 percent of UPS capacity.
3. Aggregate Load for Fluorescent Lights with Electronic Ballasts having loadcurrent rated at 15 Percent THD: 5 percent of UPS capacity.
4. Aggregate Load for Motors, 3-Phase-Induction Type, Random Across-theLine Starting: 15 percent of UPS capacity. Largest individual motor full-load
kVA is 10 percent of UPS capacity.
5. Aggregate Load for Motors, Single-Phase, Capacitor Start, Induction Run: 5
percent of UPS capacity. Motors operate continuously.
6. Aggregate Load for High-Intensity-Discharge Lighting (High-PressureSodium Type, Photoelectrically Controlled): 5 percent of UPS capacity.
7. Miscellaneous Linear Loads: 10 percent of UPS capacity.
2.6 PERFORMANCE EFFICIENCIES
A. Overall system efficiency, when operated within indicated nominal input- and
output-voltage and frequency limits, is within the following minimums (choose
the size for your application):
89
1. 30-kVA and Smaller Systems: 65 percent at 100 percent load, 60 percent at
75 percent load, and 55 percent at 50 percent load.
2. 37.5- to 74-kVA Systems: 78 percent at 100 percent load, 77 percent at 75
percent load, and 75 percent at 50 percent load.
3. 75- to 124-kVA Systems: 84 percent at 100 percent load, 83 percent at 75
percent load, and 82 percent at 50 percent load.
4. 125- to 224-kVA Systems: 88 percent at 100 percent load, 87 percent at 75
percent load, and 86 percent at 50 percent load.
5. 225-kVA and Larger Systems: 90 percent at 100 percent load, 89 percent at 75
percent load, and 88 percent at 50 percent load.
B. Maximum Acoustical Noise: 58 dB, "A" weighting, emanating from the system
under any condition of normal operation, measured 36 inches (900 mm) from the
nearest surface of the enclosure.
C. Maximum Energizing Inrush: 6 times the full-load current.
D. Maximum Output-Voltage Regulation for loads up to 50 percent unbalanced:
plus or minus 2 percent of the full range of battery voltage.
E. Output Frequency: 60 Hz, plus or minus 0.5 percent of the full range of input
voltage, load, and battery voltage.
F. Maximum Harmonic Content of Output-Voltage Waveform: 5 percent RMS total
and 3 percent RMS for any single harmonic for rated full linear load more than
the full range of battery condition and input voltage and frequency.
G. Overload Capacity of System at Rated Voltage: 125 percent of full-load rating
for 10 minutes and 150 percent for 10 seconds.
H. Maximum Output-Voltage Transient Excursions from Rated Value: For the
following instantaneous load changes, stated as percentages of rated full load,
voltage shall remain within the stated percentages of rated value and recover to
within plus or minus 2 percent of that value within 100 ms:
1.
2.
3.
4.
50 Percent: Plus or minus 8 percent.
100 Percent: Plus or minus 10 percent.
Loss of AC Input Power: Plus or minus 5 percent.
Restoration of Input Power: Plus or minus 5 percent.
I. Normal mode (Common mode) EMI noise attenuation range over 10 kHz to
50 MHz range: 60−80 dB. If this attenuation is not attainable, EMI filters at
the input and output conductors shall be installed into the UPS equipment.
90
NOTE : When used in HF and VHF radio receiving sites and sensitive
data-processing facilities, all conductors entering and exiting the UPS
(including AC, DC, and ground conductors) must meet one of the
following EMI requirements:
1. FCC Class B requirements.
2. The conducted current limitations of Section of MIL-STD- 461.
3. The conducted current limitations provided by the SNEP program.
If these limitations are not met by a standard model of an UPS, EMI
filters must be added to the input and output power conductors to
meet the stated conducted current limitations over the frequency
range of 10 kHz to 50 MHz.
2.7
SYSTEM COMPONENTS, GENERAL
A. Description: Solid-state devices using hermetically sealed semiconductor
elements. Devices include rectifier/battery charger, inverter, static bypass transfer
switch, and system controls.
B. Enclosure: Provide separate cabinets or separate compartments of enclosures for
major components such as static bypass transfer switch, rectifier, battery, inverter,
and maintenance bypass.
C. Control Assemblies: Mount on modular plug-ins, arranged for easy maintenance.
D. Surge Suppression: Protect UPS system input elements, rectifier/battery charger,
inverter, controls, and output components against voltage transients with surge
suppressors listed in UL 1449, and tested according to IEEE C62.41, Category B.
E. Power Assemblies: Mount rectifier and inverter sections and static bypass
transfer switch on modular plug-ins, arranged for easy maintenance.
F. Design and fabricate internal supports for assemblies, subassemblies,
components, supports, and fastenings for batteries to withstand static and
anticipated seismic forces in any direction, with the minimum force value used
being equal to the equipment weight.
2.8 RECTIFIER/BATTERY CHARGER
A. Capacity: Adequate to supply the inverter during full output load conditions and
simultaneously recharge the battery from fully discharged condition to 95 percent
91
of full charge within 10 times the rated discharge time for duration of supply
under battery power at full load.
B.
Input Current Distortion: Harmonic suppression, either by input harmonic
filters or inherent in the rectifier/battery charger design, reduces total harmonic
content of the current drawn from the input power source by the system to less
than 10 percent for sources with X/R ratios from 2 to 30. This applies for all UPS
load currents from 0 to 100 percent of full load.
C. Input Current Distortion: Less than 32 percent THD at rated UPS load. (change
to 10 percent THD)
D. Rectifier Control Circuits: Immune to frequency variations within the rated
frequency range of the system. Response time can be field adjusted for maximum
compatibility with local generator-set power source.
E. Battery float-charging conditions, in terms of voltage and charging current under
normal operating conditions, are within battery manufacturer's written instructions
for maximum battery life.
F. Input Power Factor: At least 0.85 lagging when supply voltage and current are at
nominal rated values and UPS are supplying rated full load.
2.9
BATTERY (Choose one type depending on installation application)
A. Description: Valve-regulated, recombinant, lead-calcium units, factory assembled
in an isolated compartment of UPS cabinet, and complete with battery disconnect
switch.
B. Description: Valve-regulated, recombinant, lead-calcium units, factory assembled
in a separate cabinet that matches UPS cabinet in appearance. Equip battery
assembly with battery disconnect switch and arrange for drawout removal of the
battery assembly from the cabinet for inspection and test.
C. Description: Lead-calcium, heavy-duty, industrial type in styrene acrylonitrile
containers mounted on 3-tier, acid-resistant, painted steel racks arranged as
indicated. Assembly includes a battery disconnect switch, intercell connectors, a
hydrometer syringe, and a thermometer with specific gravity-correction scales.
92
2.10
BATTERY-MONITORING SYSTEM
A. Battery ground-fault detector initiates an alarm when resistance to ground of
positive or negative bus of battery is less than 5000 ohms.
B. Battery compartment smoke/high-temperature detector initiates an alarm when
smoke or a temperature greater than 75 deg C occurs within the compartment.
C. Automatically measure and electronically record individual cell voltage,
impedance, and temperature, plus total battery voltage and ambient temperature.
Measure parameters on a routine schedule selected by the operator. Measure
battery and cell voltages and time to the nearest second during battery-discharging
events such as utility outages. Monitoring system includes the following:
1. Factory-wired sensing leads to cell and battery terminals and cell temperature
sensors.
2. Modem and connectors for data transmission via RS-232 link and external
signal wiring to a computer. External signal wiring and computer are not
specified in this Section.
3. Software designed to store and analyze battery data using an IBM-compatible
computer, which is not specified in this Section. Software reports individual cell and
total battery performance trends and provides data for scheduling and prioritizing
battery maintenance.
D. Automatically measure and electronically record individual cell voltage,
impedance, temperature, and electrolyte level, plus total battery voltage and
ambient temperature. Measure parameters on a routine schedule selected by the
operator. Measure battery and cell voltages and time to the nearest second during
battery-discharging events such as utility outages. Monitoring system includes
the following:
1. Modem for data transmission via RS-232 link and external signal wiring to a
computer. External signal wiring and computer are not specified in this
Section.
2. Software designed to store and analyze battery data using an IBM-compatible
computer, which is not specified in this Section. Software reports individual
cell and total battery performance trends and provides data for scheduling and
prioritizing battery maintenance.
93
2.11
INVERTER
A. Description: Pulse-width modulated, with sinusoidal output. Include a bypass
phase synchronization window to optimize compatibility with local generator-set
power source.
2.12 STATIC BYPASS TRANSFER SWITCH
A. Switch Rating: Continuous duty at rated full load. Switch provides make-beforebreak transfer. A contactor or electrically operated circuit breaker in the inverter
output provides electrical isolation.
2.13 MAINTENANCE BYPASS/ISOLATION SWITCH
A. Comply with NEMA PB 2 and UL 891.
B. Switch Rating: Continuous duty at rated full load of system.
C. Mounting Provisions: Locate inside one of the modular system cabinets, behind a
lockable door.
D. Mounting Provisions: Separate wall- or floor-mounted unit as indicated.
E. Key interlock requires unlocking maintenance bypass/isolation switch before
switching from normal position with key that is released only when UPS are
bypassed by static bypass transfer switch. Lock is designed specifically for
electrical component interlocking.
2.14 OUTPUT DISTRIBUTION SECTION
A. Panelboard: Comply with Division 16 Section "Panelboards" for panelboards
with circuit breakers and other features as indicated in a panelboard schedule.
Match and align panelboard cabinet with other UPS cabinets.
2.15
INDICATION AND CONTROL
A. General: Group displays, indications, and basic system controls on a common
control panel on the front of UPS enclosure.
B.
Minimum displays, indicating devices, and controls include those in lists below.
Provide sensors, transducers, terminals, relays, and wiring required to support
listed items. An audible signal sounds for alarms as well as the visual indication.
94
C. Indications: Labeled LED display.
D. Indications: Plain-language messages on a liquid crystal or digital LED display.
1. Quantitative Indications: Include the following:
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
Input voltage, each phase, line to line.
Input current, each phase.
Bypass input voltage, each phase, line to line.
Bypass input frequency.
System output voltage, each phase, line to line.
System output current, each phase.
System output frequency.
DC bus voltage.
Battery current and direction (charge/discharge).
Elapsed time-discharging battery.
2. Status Indications: Include the following:
a.
b.
c.
d.
e.
Normal operation.
Load on bypass.
Load on battery.
Inverter off.
Alarm condition exists.
3. Alarm Indications: Include the following:
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
Bypass AC input overvoltage or undervoltage.
Bypass AC input overfrequency or underfrequency.
Bypass AC input and inverter out of synchronization.
Bypass AC input wrong-phase rotation.
Bypass AC input single-phase condition.
Bypass AC input filter fuse blown.
Internal frequency standard in use.
Battery system alarm.
Control power failure.
Fan failure.
UPS overload.
Battery-charging control faulty.
Input overvoltage or undervoltage.
Input transformer over temperature.
Input circuit breaker tripped.
Input wrong-phase rotation.
Input single-phase condition.
95
r.
s.
t.
u.
v.
w.
x.
y.
z.
aa.
ab.
ac.
ad.
ae.
af.
Approaching end of battery operation.
Battery undervoltage shutdown.
Maximum battery voltage.
Inverter fuse blown.
Inverter transformer over temperature.
Inverter over temperature.
Static bypass transfer switch over temperature.
Inverter power-supply fault.
Inverter transistors out of saturation.
Identification of faulty inverter section/leg.
Inverter output overvoltage or undervoltage.
UPS overload shutdown.
Inverter current sensor fault.
Inverter output contactor open.
Inverter current limit.
4. Controls: Include the following:
a.
b.
c.
d.
e.
Inverter on-off.
UPS start.
Battery test.
Alarm silence/reset.
Output-voltage adjustment.
E. Analog Meters: Accurate within 2 percent.
F. Dry Form "C" Contacts: Available for remote indication of the following
conditions:
1.
2.
3.
4.
G.
UPS on battery.
UPS on-line.
UPS load on bypass.
UPS in alarm condition.
Remote Status and Alarm Panel: Labeled LEDs indicate conditions listed
above. Audible signal indicates alarm conditions. Silencing switch in face of
panel silences signal without altering visual indication.
1. Cabinet and Faceplate: Surface- or flush-mounted to suit mounting conditions
indicated.
2.16 REMOTE UPS CONTROL AND MONITORING SYSTEM
A. Description: A remote microprocessor for the unit control panel to indicate
alarms and to control as specified in "Indication and Control" Article above.
Record power-line transients and provide analytical capability. Include the items
96
described below, but do not include the remote computer or the connecting signal
wiring. System includes the following:
1. Modem and connectors for data transmission via RS-232 link and external
signal wiring to a computer. External signal wiring and computer are not
specified in this Section.
2. Software designed to secure control and monitoring of UPS functions and to
provide on-screen explanations, interpretations, and action guidance for
monitoring indications. Include on-screen descriptions of control functions
and instructions for their use. Permit storage and analysis of power-line
transient records. Design for an IBM-compatible computer, which is not
specified in this Section.
2.17 MECHANICAL FEATURES
A.
Enclosures: NEMA 250, Type 1.
B. Ventilation: Redundant fans or blowers draw in ambient air near the bottom of
the cabinet and discharge it near the top rear.
2.18 SOURCE QUALITY CONTROL
A. Factory test complete UPS, including battery, before shipment. Include the
following tests:
1. Functional test and demonstration of all functions, controls, indicators, sensors,
and protective devices.
2. Full-load test.
3. Transient-load response test.
4. Overload test.
5. Power failure test.
6. Efficiency test at 50, 75, and 100 percent loads.
B. Observation of Test: Give 14 days advance notice of tests and opportunity for
Government's representative to observe tests.
C. Report test results. Include the following data:
1. Description of input source and output loads to be used. Describe actions
required to simulate source load variation and various operating conditions
and malfunctions.
97
2. List of indications, parameter values, and system responses considered
satisfactory for each test action. Include tabulation of actual observations
during test.
3. List of instruments and equipment required to duplicate factory tests in the
field for those tests required to be repeated there.
PART 3 - EXECUTION
3.1 INSTALLATION
A. Install system components on 4-inch- (100-mm-) high concrete housekeeping
bases. Cast-in-place concrete, reinforcing, and formwork are specified in
Division 3.
B. Maintain minimum workspace at equipment according to manufacturer's written
instructions and NFPA 70.
C. Connections: Interconnect system components. Make connections to supply and
load circuits according to manufacturer's wiring diagrams, unless otherwise
indicated.
3.2 IDENTIFICATION
A. Identify components according to Division 16 Section "Electrical Identification."
1. Identify each battery cell individually.
3.3 FIELD QUALITY CONTROL
A. Manufacturer's Field Service: Supervision of unit installation, connections, tests,
and adjustments by a factory-authorized service representative. Report results in
writing.
B. Manufacturer's Field Service: Supervision of unit installation, connections,
pretests, and adjustments by a factory-authorized service representative. Report
results in writing.
C. Supervised Adjusting and Pretesting: Under supervision of a factory-authorized
service representative, pretest system functions, operations, and protective
features. Adjust to ensure operation complies with specifications. Load the
system using a variable-load bank simulating kVA, kW, and power factor of loads
for which unit is rated.
98
D. Tests: Perform tests listed below by an independent testing agency meeting the
qualifications specified in the "Quality Assurance" Article. Perform tests
according to the manufacturer's written instructions. Load the system using a
variable-load bank to simulate kVA, kW, and power factor of loads for the unit's
rating. Use instruments calibrated, within the previous 6 months, according to
NIST standards.
1. Simulate malfunctions to verify protective device operation.
2. Test duration of supply on emergency, low-battery voltage shutdown, and
transfers and restoration due to normal source failure.
3. Test harmonic content of input and output current less than 25, 50, and 100
percent of rated loads.
4. Test output voltage under specified transient-load conditions.
5. Test efficiency at 50, 75, and 100 percent rated loads.
6. Test remote status and alarm panel functions.
7. Test battery-monitoring system functions.
E. Retest: Correct deficiencies and retest until specified requirements are met.
3.4 CLEANING
A. On completion of installation, inspect system components. Remove paint
splatters and other spots, dirt, and debris. Repair scratches and mars of finish to
match original finish. Clean components internally using methods and materials
recommended by manufacturer.
3.5 DEMONSTRATION
A. Engage a factory-authorized service representative to train Government's
maintenance personnel as specified below:
1. Train Government's maintenance personnel on procedures and schedules
related to startup and shutdown, troubleshooting, servicing, and preventive
maintenance.
2. Review data in the operation and maintenance manuals. Refer to Division 1
Section "Contract Closeout."
3. Review data in the operation and maintenance manuals. Refer to Division 1
Section "Operation and Maintenance Data."
4. Schedule training with Government, through COR, with at least 7 days
advance notice.
99
3.6 COMMISSIONING
A. Battery Equalization: Equalize charging of battery cells according to
manufacturer's written instructions. Record individual cell voltages.
100
LIST OF REFERENCES
1. IEEE Std 473-1985, IEEE Recommended Practice for an Electromagnetic Site Survey
(10 kHz to 10 GHz), IEEE, Inc., 345 East 47th Street, New York, NY, 1985.
2. E.F. Vance, “Electromagnetic Interference Control”, IEEE Trans. on Electromagnetic
Compatibility, pp. 319-328, EMC 22, November 1980.
3. W. Graf and E.F. Vance, Topological Approach to the Unification of Electromagnetic
Specifications and Standards, Proceedings of the IEEE National Aerospace and
Electronics Conference, Dayton, Ohio, May 1982.
4. E.F. Vance, W. Graf, and J.E. Nanevich, Unification of Electromagnetic Specifications
and Standards, Part 1: Evaluation of Existing Practices, DNA Report 5433F-1,
Prepared by SRI International, 31 October 1980.
5. W. Graf, J.M. Hamm, and E.F. Vance, Unification of Electromagnetic Specifications
and Standards, Part II: Recommendations for Revisions of Existing Practices, DNA
Report 5433F-2, Prepared by SRI International, 28 February 1983 (Revised 15 April
1983).
6. Vincent, W. R. and Munsch, G. F., Signal-to-Noise Enhancement Program PowerLine Noise Mitigation Handbook, Second Edition, prepared for Headquarters, Naval
Security Group, Washington, DC, January 1993.
7. Technical Product Catalog, Delta High Performance Filters, DR Series, Delta Inc.,
Fremont, CA, pp. 6-8.
8. EEM 2001, Electronic Engineer’s Master, (www.eem2001.com), Hearst Business
Communications, Garden City, NY, 2001.
101
THIS PAGE INTENTIONALLY LEFT BLANK
102
BIBLIOGRAPHY
O’Dwyer, J., M., “Electromagnetic Noise and Interference at High Frequency
Communications Receiver Facilities”, Master’s Thesis, Naval Postgraduate School,
Monterey, California, June 1984.
Vincent, W. R., Briefing Notes, “Comments about MIL-HDBK-419A: Grounding,
Bonding, and Shielding for Electronic Equipment and Facilities", unpublished.
Vincent, W. R., Adler R. W., and Parker A. A., “The EMI Aspects of Grounds At
Receiving and Data-Processing Facilities”, Technical Memorandum, SNEP 9803, March
1998
Eckert, J.K., A Handbook Series on Electromagnetic Interference and Compatibility,
Volume 9, Commercial EMC Standards of the United States, Interference Control
Technologies, Inc., Gainesville, Virginia, 1988.
Wintzer, C.M., A Handbook Series on Electromagnetic Interference and Compatibility,
Volume 10, International Commercial EMC Standards, Interference Control
Technologies, Inc., Gainesville, Virginia, 1988.
Hodge, J. W., "A Comparison Between Power Line Noise Level Field Measurements and
Man-Made Radio Noise Prediction Curves in the High Frequency Radio Band", Master's
Thesis, Naval Postgraduate School, Monterey, California, December 1995.
Equipment User Manual, FERRUPS FE/QFE 500 VA – 18 kVA UPS Technical Manual,
Best Power Technology, Inc., Necedah, Wisconsin, 1996.
IEEE Std 446-1995 , Recommended Practice for Emergency and Standby Power Systems
for Industrial and Commercial Applications, ANSI, IEEE Orange Book
IEEE Std 1100-1999 , Recommended Practice for Powering and Grounding Sensitive
Electronic Equipment, ANSI, IEEE Emerald Book
MIL-HDBK-419, Grounding, Bonding, and Shielding for Electronic Equipment and
Facilities, Volume 1 (Basic Theory) and Volume 2 (Applications).
103
National Fire Protection Association (NFPA) 70, National Electric Code, 1999 edition
National Fire Protection Association (NFPA) 75, Standard for the Protection of Electric
Computer / Data Processing Equipment, 1999 edition
104
INITIAL DISTRIBUTION LIST
1. Defense Technical Information Center
Ft. Belvoir, VA
2. Dudley Knox Library
Naval Postgraduate School
Monterey, CA
3. Chairman, Code EC
Department of Electrical and Computer Engineering
Naval Postgraduate School
Monterey, CA
4. Prof. Richard W. Adler
Department of Electrical and Computer Engineering
Naval Postgraduate School
Monterey, CA
5. Prof. Wilbur R. Vincent
Department of Electrical and Computer Engineering
Naval Postgraduate School
Monterey, CA
6. Mr. Andrew A. Parker
Department of Electrical and Computer Engineering
Naval Postgraduate School
Monterey, CA
7. Edward Beran, V41
Department of Defense
8 Prof. Vincente Garcia
New Mexico State University
Department of Electrical Engineering
105
9. USAINSCOM
IALO-E
ATTN: Ms. Anne Bilgihan
10. COMNAVSECGRU
N-31
ATTN: Mr. Frank Cawley
106
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement