842r EG/ &TB w7-11

.......... 0~

TECHNICAL REPT

1713,L4,

~III~EECTROMAGNETIC PULSE (EMP) HANDBOOK

I FOR

AIR FORCE COMMUNICATIONS SERVICE

COMMUNICATIONS-ELECTRONICS-METEOROLOGICAL ENGINEERSo

11*

Sdistribution is unlimited.

SUIPPORT ENGINEERING BRANCH18.2 ELECTRONICS ENGINEERING GROUP (EEG)

RICHARDS-GEBAUR AFB, MO, 64030 /9 g

29 OC ..ER 1976r'~1(~i i<3

° !,

1842 ELECTRONICS ENGINEERING GROUP

MISSION

The 1842 Electronics Engineering Group (EEG) isorganized as an independent group reportingdirectly to the Commander, Air Force Communica-tions Service (AFCS) with the mission to providecommunications-electronics-meteorological (CEM)systems engineering and consultive engineeringfor AFCS. In this respect, 1842 EEG responsi-bilities include: Developing engineering andinstallation standards for use in planning,programming, procuring, engineering, installingand testing CEM systems, facilities and equip-ment; performance of systems engineering of CEMrequirements that must operate as a system orin a system environment; operation of a special-ized Digital Network System Facility to analyzeand evaluate new digital technology for applica-tion to the Defense Communications System (DCS)and other special purpose systems; operation ofa facility to prototype systems and equipmentconfigurations to check out and validate engi-neering-installation standards and new installa-tion techniques; providing consultive CEMengineering assistance to HQ AFCS, AFCS Areas,MAJCOMS, DOD and other government agencies.

fECu#1ITY Cl. ',I1*V!CATIO.'N OP- THI, qr {yw'~,,t fltAg , ,,t.,.,,

REP T Docu NTATON PAGE READ INSTRUCTION1DM AE { flPORr. COMPtI.riNc FORMvR EPORf ,UKI " 1 GOVT AC NO.S RN -N'"T" CATALOG NJU-.,rr

1842 EE GIEPEUE TR r7- // - I ...4. TITLE (and Su itt.) Ta , ,.tPrjTrA P9,41oo COVERED

EMP Handbook for AFCS C-E-M Engineers N/AS 6ERFOMING OG. REPORT NUM8ER

7. AUTHOR(u) I e. CON IAC YO.7 GR'MNt NUMER()

John F. Zych I N/AS. PERFORMING ORGANIZATION NAME AND ADDRESS PCGPA.. MLEMENT. PROJECT. TASK

~-~~ ,A WORK UNIT NUMOERS

1842 EE Gp/EPEUERichards-Gebaur AFB MO 64030 N/A

11. CONTROLLING OFFICE NAME AND ADDRESS It. REPORT OAYR

1842 EE Gp/EPEUE II_ Nov 76Richards-Gebaur AFB MO 64030 IS. NJMBER nP PAGES

14. MONITORING AGENCY NAME A ADDRIESS(il difterent from Controlling Offlce) IS. SECURITY CL.ASS. (of tid repotf)

1842 EE Gp/EPEUIRichards-Gebaur AFB MO 64030 Unclassifted

IS. OECLASSIlICATION/OOWNGRADINGSCHEOULE

16. DISTRIBUTION STATEMENT (of this Repor) b

--- "-, /or public relecwe and sale, il-

distribution is unlirmjted.

17. OISTRIBUTION STATEMENT (of the abstract entered In Block 20. It different from Report)

IS. SUPPLEMENTARY NOTES

N/A

It. KEY WORDS (Continue on rereree aide If neceeary and identity by block number)

EMP, Shielding, Cabling, EMP Protection

20. ABSTRACT (Continue on reverse side It necessary and Identify by block nuber)

4"This document is to provide basic engineering information about D4P. Emphasisin this document vil be on EMP protection guidelines. Information presentedis designed 0o assist engineers in familiarizing themselves with 41' and 1qPProblems.

f O. 1473 EDITION OF INoV,, OUSOLRT%

APPROVAL PAGE

This report has been reviewed and in approved for publication and dis-

tribution.

GRL T. HARISChief, Electronics G Base Systems Engrg Division

Chief, Support Engineerng Branch

NTiS YWhite Ion~j¢

DDCNOC Buff Semton

REi Classified reference# distribution

unlimited-Ho change per Kr. John F. Zych, 1842

F-EG EEIS 0

EMP HANDBOOKTABLE OF CONTENTS

A. INTRODUCTION . . . .. . . .. . . . . . . .. .. Pae

B. PURPOSE . . . . . . . . .. . . . ............ . . .1

C.. EMPsUENiRATION AND EFFECTS ............ 31. EMP Environment . . . . . . . . . . . . . . . . . 3

, 1.1 EMP Phenomenon . . . . . . . . . ..... . 3

1.2 Nuclear Weapon Effects . . . . . . . . . . . . . 31.2,1 Source Region . . . . . . . . . . . . . . . . . 31.2.2 Fields Beyond Source Region .......... 71.3 Comparison with Lightning . . .......... 7

2. Basic Atomic and Nuclear Physics . . . . . . . . . 9

3. Gamma Rays Cause EMP . . . . . . . . . . . . . . . 10

4. Typical High Altitude Burst Example . . . . . . . 11

S. Summary of EMP Generation. . . . . . . . . . . . . 13

b. EMP Effects .. . . . . . . . . . . . . . 136.1 Electric Field'Effects' is ...... 16.2 EMP Effects Below Ground . . . . . . . . . . . . 16

D. EMP PROTECTION . .. . . . .. . . .. .. . . .. 17

1. Introduction . . . . . . . . . . . . . . . . . . 172. liMP Protection Philosophy (General) . . . . . . 172.1 Application to Existing Facilities (General). . . 203. Energy Collection . . . . . . . . . . . . . . . . 234. Response of Components . . . . .............. . 23S. Design Practices . . . . . . . . . . . . . . . .. 27

5.1 System Aspects . . . . . . . . . . . . . . . . 275.1.1 System Level Concept and Definition . . . . . . 27S.1.1.1 Zoning . . . . . . . . . . . . . . . . . . . 285.1.1.2 Clustering . . . 305.1.1.3 Layering . . . . .......... . . . 305.1.1.4 Ringing . . . . . . .. . . . . . . . . .. . 305.1.1.5 Damping . . . . . . . . . . .. . . .. .. . 315.1.1.6 Violation andFixes . . . . . . . . . . . . . 315.1.1.6.1 Circuit Considerations . . . . . . . . . . 315.2 Shielding . . . . . . . . . . . . . . . . . . . . 325.2.1 1iow Do UMP Shields Work? . . . . . . . . . . . 325.2.2 Ideal Versus Realistic Shielding .. ..... 325.2.3 Shielding Effectiveness . . . . . . ...... 32

1I .... ' .. ......... ... . ... . . [ I ll I I | 11 I [ I [I

Page5.2.3.1 Absorption Loss. . . . . .......... .5.2.3.2 Reflection Loss. . . . . . . . . .5.2.4 A Shield is a Magnetic Field Reducer . . ....

5.2.5 Diffusion Shielding . . . . ....... . . "5.2.6 Apertures. . . . . . . . . . . . . . . . . . 395.2.6.1 Seams . . . . . . . .. . . . . . . . . . . . 95.2.b.1.1 Dilemmas . . . . . . . . . .. . . . . 405.2.7 Gaskets and Bonds. . . . . . . . . . 405.2.7.1 Seam Bonds . . . . . . . ............ 405.2.7.2 Clean Contacts . . . . . . . . . . . . . 405.2.7.3 Pressure Contacts ......... 415.2.7.4 Electromagnetic Gaskets and Panels . ..... 415.2.8 Shielded Enclosures ..... . . . . . . . 415.2.8.1 Finger Stock and Doors ...... . ..... 425.2.8.2 Protection Maintenance ..... . . . . . 425.2.90 Open Aperture ..... . . . . . . . . . . 425.2.10 Screens . . . . . . . . . ......... . 435.2.11 Waveguide Schemes ..... . . . . . . . . 435.3 Penetrations . . . . . . . . . . . . . . . . .. 435.3.1 Electrical Penetrations ..... . . . . . 435.3.2 Mechanical Penetrations ..... . . . . . 445.4 Cables . . . . . . . . . . . . . . . . . . . . 445.4.1 What is a Cable? . . . . . ......... . 445.4.1.1 Cable System Design . ......... . 445.4.1.2 System Configuration Options..............465.4.1.3 High Operating Levels . . . . . . 465.4.1.3.1 Component Distribution ..... . . . . 465.4.1.4 Carrier Systems . . . . . 465.4.1.5 Circuit Arrangement . . . . . . . 475.4.1.6 The Zoning Role ..... . . . . . . . . 475.4.2 Cable Fabrication ............... 475.4.2.1 Cable Types . . . . . . . . . . . . . . . . . 475.4.2.2 Why is the "Exterior" So Important? . . . . . 475.4.2.3 Braid Transparency . . . . . . . . . . . . . . 485.4.2.4 Outer Shield Construction . . . . . . . . . . 485.4.2.5 Quality Control . . . . . . . . . . . . . . . 495.4.2.6 Outer Jacket or Sheath. . . . . . . . . . . . 49S.4.3 Damping Schemes. . . . . . . . . . . . . . . . . 495.4.4 Cable Terminal Treatment. . . . . . . . . . . . 50

o.4.5 PlugsandSockets ............... 505.5 Conduits . . . . . . . . . . . . . . . . . . . . . So

5.6 End Boxes . . . s55.7 Internal Treatment ia . . . . . Si5.7.1 Twisting and Shielding . . . . . . . . . . . . . S15.7.2 Balanced Isolation . . . . . . . . . . . . . . . S15.7.3 Terminations . . . . . . S 15.7.4. Power Supply and Control Circuits 525.7.5 Hierarchies and Zone Loss ........... 52

.- :-- .v.- -r4--*-'- - -

Page5.8 Grounding............ ... . . . . . .. 525.8.1 Ground Semantics * 25.8.2 Why is liMP Concerned'with Outside Grounds?, S 35.8.3 Ground Concept . . .. . . *. . . . . .* . . S3

5.8.4 Ground Quality . . ... . . . . . . . . . . . 545.8.4.1 Exterior Ground Resistance and Transient Impedance S5.8.5 Inside Grounds . ........... 55s.8.S.1 Interior Grounding Systems .. ......... 555.8.5.2 Internal Grounds ............ . . . . . 555.9 Bonds . .............. . . . . . . . . . . . 575.9.1 Interzonal Bonds . . . . . . . . . . . . . . . . 75.9.1.1 Bond Straps . . . . . . . . . . . . . . . . .. . 575.9.2 Good Bond Strap Practice .. . . . . . .. . . . . s75.10 Protective Devices and Techniques . . . ........ 575.10.1 What is a Protective Device? . . . . . . . . . . . 595.10.2 Basic Protection Element Concept . . . . . . . . . 595.10.3 Categorization. . . . . . . . .......... . 595.10.4 Circuit Isolation . . . . . . . . . . . . . . . 605.10.5 Inductive Devices . . . . . . . ....... .. 605.10.6 Passive Devices - Filters . . . . . . ........ 605.10.7 Butterworth Filters ............... 605.10.8 Device Construction . . . . .............. 605.10.9 Device Installation ....... ............... 615.10.10 The EMP Box . ...... ..... . . . . . . 615.10.11 Active or Non-linear Devices ....... . .. 615.10.12 Hybrids . . . . . . .. ........... 625.10.13 Types of Non-linear Devices 625.10.13.1 Consequences of Non-linear Operation .... . 625.10.13.2 Spark Gaps and Gas Diodes ............. 6 35.10.13.3 Zener and Silicon Diodes ........... 635.10.13.4 Thyrites . . . . . . . . . . . . . . . . . . . 635.10.14 Fast Relays. . . . . . . .............. 6 35.10.15 Crowbar Circuits . . . .. . . . . . . . . . . . 6 35.11 Circumvention Techniques . . . . . . . . . .... 645.11.1 System Constraints .. . . . . . . . . . . 64S.11.2 Non-threat Specific Schemes. . . ........... ... 6 55.11.3 Threat Specific Schemes ........... .... 655.11.4 Non-electrical Schemes . . ............. 6 55.12 Economics . . .. . . . . ... . . . . . . . . . . 6 5

iii -

LIST OF ILLUSTRATIONS

Figure No. Caption PageI Area of Coverage of EMP from High Altitude Detona-

2 Energy Floiaga .' ; .; g ; a . .. 43 Surface Burst.... . . . . . . . . . . . S4 Illustration of the Basic Geometry of the Hgh

Altitude Burst . . . . . . . . . . . . . . . . 6S EMP fromlli~h Altitude Bursts . . . . . . . . . . 126 Tangent Radius for a High Altitude Burst . . . . 147 Typical Mapping of Zoning Approach . . . . . . . 298 Typical Shielding Design Problems .......... . . 339 Typical Ring Ground System . . . . . . . . . . . 56

10 Typical Bond Strap. . . . . . . . . . . . . . . . 58

ei

iv

EMP HANDBOOK

A. INTRODUCTION. An area that very few C-E-NI engineersare aware of is the effects on our communication systemsin the event of a nearby nuclear explosion. One effectthat was found is that under the proper circumstances, asignificant portion of the energy released during a nucleardetonation can be made to appear as an Electromagnetic Pulse(EMP) having the same frequencies as our commercial radioand military communication systems. This document willinvestigate this phenomenon and provide typical engineeringinformation that must be considered to provide adequateprotection from EMP in any military communication system.The engineering information to be provided is, in mostcases, familiar terminology and closely ties into EMC/EMIprotective procedures. Engineering principles are similarto ones in use today ant will be most familiar to engineerswith EMI/EMC experience. One of the primary reasons whyDIP has great interest today is the fact that EMP iscapable of disabling electrical and electronic systems asfar as 3000 miles from the site of the detonation. Thismeans that a high yield nuclear weapon burst above theatmosphere could be used to knock out improperly designedelectrical and electronic systems over a large area of theearth's surface without doing any other significant damage.Figure 1 shows this situation.

Another point of concern in EMP is the strong electro-magnetic field created. An idea of the amplitude of theEMP electromagnetic field can be gained when compared withfields from man-made conventional sources. A typical highlevel IP pulse could have an intensity of 100,000 voltser meter. This is 1,000 times more intense than a radaream of sufficient power to cause biological damage such

as blindness or sterilization.

B. PURPOSE. This document is to provide basic engineeringinformation about EMP that is presently available. EMP isa large and complex subject and this document will notdelve into involved mathematical investigations but generalinformation. Emphasis in this document will be basic EMPprotection guidelines. The information presented is designedto assist engineers in familiarizing themselves with DMP andEMP problems. This handbook will be basically a compilationof existing data on EMP and only portions that may be ofinterest or of importance are included. A comprehensivebibliography will be included for further reference.

I " I ll llll I II I , ... ... i ,1

z a.0w

w

C. ELECTROMAGNETIC PULSEGENERATION ix\,D EFFECTS

1. EMP ENVIRONMENT DESCRIPTION. A nuclear detonation generateslarge amounts of energy which can be grouped in such cate-gories as blast, thermal radiation, nuclear radiation, andelectromagnetic pulse (EMP). This chapter provides anintroduction to EMP generation and effects.

1.1 EMP Phenomenon. A fundamental phenomenon existswhich creates the electromagnetic pulse. The basic mecha-nisn is electron scattering by the collision of gamma rayswith air molecules or other materials. The collisionknocks electrons free of the air molecules and causes theelectrons to move rapidly away from the center of theexplosion and from the now positively charged parent airmolecules. This separation of charges, occurring on awholesale basis, creates electromagnetic fields. An energyflow diagram which illustrates the transformation of energyinvolved in the process of EMP generation is shown in Figure2. The energy released from a nuclear burst in the form ofgamma ray photons interacts with the earth's atmosphere toproduce electrons and nositive ions. The separation ofelectrons and positive ions produces an electric field. Theflow of electrons constitutes a current which radiateselectromagnetic energy, providing some asymmetry exists.

The energy contained in EMP is similar to that in EMwaves generated by a lightning strike, but the high fre-quency energy content in EMP is a much larger fraction ofthe total pulse energy.

1.2 NUCLEAR WEAPON EFFECTS. The relative importance ofall nuclear weapons effects, including EMP, depends onweapon characteristics, burst point, and position of thesystem of interest. Emphasis in this document will be theEMP fields generated by a high altitude burst and some dis-cussion about surface bursts.

1.2.1 Source Region. For both the high altitude and sur-face bursts, intense fields appear in what is called thesource region. The source region for a surface burst islimited to about a two to ten kilometer diameter about theburst. Figure 3 illustrates a surface burst. For a highaltitude burst, the source region can be on the order of3000 kilometers in diameter. Figure 4 illustrates a h gh

3

hmI

I

clcJ

z e0 zAJ

F w:E U)

4 hi

zz0-

IL

f.-. '0 _ 1

0!; 0

I cc

WuI

v a%^IIIIg

6

altitude burst. This source region extends from about 20to 40 kilometers in altitude. In the source region, weaponeffects other than ENP must also be considered. As noted,the size of the source region is severely confined by theatmosphere for an atmospheric burst. However, the pressurepulse, which physically damai.es the buildings, is notsimilarly restricted. Thus, in the case of soft systems(buildings), the most severe lIMP exposure at otherwisesurvivable points is associated with the low-yield bursts.

1.2.2 Fields Beyond Source Region. Somewhat less intensefields exist beyond the source regions. Additional mecha-nisms exist to radiate the energy of the source fields wellbeyond the source regions. In the case of a near-surfaceburst, the net charge separation caused by asymmetry of thesource reyion contributes to the more distant radiatedfields. tn an outside-the-atmosphere burst, the earth'siiiagnetic field bends the scattered electron current movingaway from the burst point. This bending produces an efficientconversion of the energy of the moving electrons into aradiated electromagnetic pulse in the radio spectrum. Thisradiation is propagated from the source region onto thesurface of the earth. In the case of a high-altitudeburst, a significant overpressure pulse does not exist nearthe surface of the earth. Almost all of the other promptweapon effects are diminished by the atmosphere, so thatthe most significant prompt weapon effect is the EMP. Asnoted previously, the source region can be quite large,in the order of 1,000 miles in diameter. As a consequence,the radiated fields from this source region can cover asubstantial fraction of the earth's surface.

1.3 Comparison with Lightning. One method of assessingtle impact of this electromagnetic pulse from a nucleardetonation is to compare the phenomenology of a nuclearexplosion to that of a lightning strike. In the case ofboth the lightning strike and the nuclear detonation, onlya fraction of the total energy is released in the form ofelectromagnetic radiation. The total energy radiated froma large nuclear detonation, however, can be many orders ofmagnitude greater than the energy radiated from a lightningstrike. Thus, while the most familiar result of theelectromagnetic radiation from a lightning strike is radio-static, the results from a nuclear burst would rvt onlycause static but could be capable of damaging se,,sitiveelectronic components. In terms of waveshapes, it is notpossible to draw explicit comparisons. In the case of

7

lightning strikes, the rise times vary widely; and bynormal time intervals, little time elapses before fullfield intensities are reached (microseconds). However,many of the rise times associated with nuclear eventscan occur in much shorter intervals (nanoseconds). Itcan be assumed that these very fast rise times can alsooccur for'nuclear-created electromagnetic environments,although wide variations from this are possible. Thus,the amplitudes and waveshapes of the EMP can be considerablydifferent from those of lightning. Another importantdifference is the spatial distribution of the electro-magnetic environment. In the case of EMP, this environmentis widely distributed; however, in the case of lightningstrikes, the most severe effects are quite localized.An additional important consideration is the timing ofthe appearance of these high-energy environments. Inthe case of EMP, this high-energy environment occursnearly simultaneously over large areas (the only limitationbeing the speed of light). On the other hand, in thecase of lightning strikes, these high-energy environ-ments seldom appear simultaneously over wide regions.The amplitude of the EMP can also be expected to exceedthe normal electromagnetic environment created bynearby broadcast stations. Other obvious differencesexist here, since the broadcast station radiation ismore or less continuous--whereas the EMP occurs as ashort duration burst of energy. The above discussionpoints up the major differences that exist betweennormally occurring electromagnetic environments and theelectromagnetic pulse created by a nuclear detonation.For certain situations, the amplitude of the EMP canexceed by several orders of magnitude, this "normal"environment. The most severe EMP environment for ahardened complex, such as a carefully hardened shelter,can occur within the ionized sphere or source region.The radiated environments at a distance are obviouslyless severe than those found in the source region. Theradiated EMP environment is of significance, however,since it can appear and must be considered over large geo-graphical areas. In Section C, a brief explanation of EMPphenomenon was discussed; however, to get a better feelof EMP, it is necessary to go back to basic atomic physicsand investigate EMP phenomenon from this standpoint. Thissection will attempt to explain from an atomic physicsstandpoint EMP, what causes EMP, and the magnitude ofEMP pulse.

8

2.1 BASIC ATOMIC AND NUCLEAR PHYSICS. The structure ofan atom can be visualized in the familiar form of asmall, positively charged nucleus surrounded by anelectron cloud. The electron cloud is held in place bythe electric coulomb attraction between the n leus andelectron. The nucleus is on the order of 10- cm indiameter, while the electron cloud is about four ordersof magnitude larger. The nucleus is made up of protons(single positivdand neutrons (zero charge). The nucleusis held together by intense, short-range forces which arenot yet coipletely understood. These nuclear forces areso strong that they overbalance the electric coulomb repul-sion of the protons for each other. Atoms and nucleiexist in states having certain discrete energies. Thisis the basis for the quantum theory developed by Bohr, Planck,and many others. This theory gives a complete explanationof chemistry and atomic physics. There can be no doubtof its essential correctness. The energy difference ofelectron levels is of the order of a few electron volts.Aj electron can fall from one level to a lower one, at thesame time emitting a quantum of light (protons. The energylost by the electron is carried away by the light quantum.For example, "green" light quanta have energy of about 2.5electron volts. An electron volt is the kinetic energygained by an electron when it is accelerated through apotential of one volt. The energy difference of protonlevels in a nucleus is of the order of a few millionelectron volts. A proton can fall from one level to alower one, at the same time emitting a gamma ray (photon).Again, energy is conserved. There is no difference betweengalx.ta rays and light quanta except that gamma rays haveabout a million times more energy per quantum. Gamma raysare the principal cause of EMP. About 0.1% of the energyof a typical nuclear bomb appears as prompt gamma rays.flow does this happen? In order to answer this question,the fission process must be investigated.

To start with, consider the action of protons. Theprotons in a nucleus repel each other electrically. Ina spherical nucleus, the electric repulsion is overbalancedby the nuclear forces. Ordinary nuclei are held sphericalby a surface tension. In large nuclei, the surface tensionis not srong enough to keep the nucleus spherical. Theelectric repulsion tends to make the nucleus elongate,eventually dividing into two roughly equal parts. Thisis the fission process. This is one of the basic reasonswhy nuclei larger than uranium do not exist in nature; they

9

are unstable against fission. Some nuclei, like L2 3 5 , arejust on the verge of being unstable. If such a nucleusis hit by a free neutron, it may undergo fission. Thefission process is not neat and tidy as a few free neutronsglislost in the rush. These freed neutrons may hit otherU nuclei, causing them to fission. On the average,a, ee neutron travels about 10 cm before striking anojherU' nucleus and making it fission. If the piece of Uis too small (sub-critical), the freed neutrons willejpe without causing further fissions. If the piece ofU is large (super-critical), the number of fissions willgrow exponentially/with time, with the number of fissionsproportional to e ". The enfolding time, T, is approx-irmately equal to l travel time of a freed neutron beforehitting another U nucleus. This time is in the orderof 10 nanoseconds. The EMP can have comparable rise time.Where do gamma rays come from? The fission fragments areusually not born in their ground levels. Free neutronscbllide witli'other nuclei in the bomb or air or earth,knocking some of these into levels above the ground level.Gaiumna rays are then emitted in transitions back to theground level.

3. GAMMA RAYS CAUSE IB3P. How do the gamma rays cause EMP?The answer is through the Compton scattering process.Compton discovered that photons can collide with electrons,knocking them out of the atoms in which they were originallybound. These Compton collisions are somewhat like thecollision of a moving billiard ball with one at rest. Therecoil electron, like the ball originally at rest, goespredom~tinantly forward after the collision. Thus, a directedflux of gamma rays produces, by Compton collisions, adirected flux of electrons. This constitutes an electriccurrent, which generates the EiMP. In order for a system ofradial currents to radiate electromagnetic energy, a depar-ture from spherical symmetry is required. Anisotropy ofthe emission of gamma rays from a burst is small and ofshort duration compared with other factors. The presence ofthe earth-atmosphere interface provides the asymmetry forsurface bursts. For high-altitude bursts, the asymmetryfactor is introduced by the atmospheric density gradientand geomagnetic field. We can now consider some order ofmagnitude estimates of the energy involved at each stage ofthfLMP generation process. A one megaton bomb releases 4 x1012 joules of energy. About 0.1% of this energy, or 4 x10 joules, may appear as prompt gamma rays. This amountof energy is equivalent to that produced by a hundred

10

maegawatt power plant running for about 11 hours. A fairfraction, about one-half of this goes into the Comptonrecoil current. Fortunately, most of this energy goes intoheating air rather than into the 1INP. About 10 -6of thegamma energy goes into EMP; thus giving about 10" of thebomb energy going into the IP.

4. TYPICAL iIGH ALTITUDE BUkST EXAMIPLE. The geometry for ahigh altitude burst is illustrated in Figure 5. In thisillustration, the height of the burst is taken as 400kilomaeters or 250 miles. For this height of burst, thedistance to tile horizon is 2250 kilometers, or 1400 miles.The resulting IMP can cover a similar area. This empha-sizes the significant aspect of an EMP from a high altitudeburst that the large amplitude fields can cover large geo-graphical regions. The outgoing gammas from the burst forma spherical shell which expands with the velocity of light.Since most of the gammas are emitted in about 10 nanoseconds,the thickness of the shell at any instant is a few meters.When the gamma shell begins to intersect the absorbinglayer of the atmosphere, an outgoing electromagnetic pulseis generated. This pulse moves along with the remaininggammas. Above about 40 kilometers altitude, the atmosphericdensity is sufficiently small that the high energy gammasare not affected appreciably. The atmospheric density islarge enough that the gammas are absorbed by Comptonscattering below 40 kilometers. The gamma absorption isnearly complete by the time they reach 20 kilometers alti-tude. The source region for a high altitude burst is thusbetween about 20 to 40 kilometers, which is approximately65,000 to 130,000 feet. At the altitude of the sourceregion, the stopping range of Compton recoil electrons is ofthe order of 100 meters. In traveling this distance, theCompton electrons are strongly deflected by the geomagneticfield with a gyro radius of about 100 meters. The Comptonrecoil current, therefore, has strong components in direc-tions transverse to the gamma propagation direction. Thistransverse current ra.diates an electromagnetic wave thatpropagates in the forward direction. The outgoing wavekeeps up with the gamma shell and is continually augmentedby the transverse Compton current until the gammas are allabsorbed. Then the electromagnetic wave goes on alone as afree wave or pulse. Secondary electrons produced by theComptons make the air conducting. This conductivity atten-uates the electromagnetic pulse. The amplitude of EMP isdetermined by a balance between:

Sk-

w

EY

CD / c0w cc

'I'I U.

0 iiwi 4iw>

aI

gI 0

(OOx

AC "a

S212

a. Increase due to transverse Compton current.

b. Attenuation due to conductivity.

S. Sl ,ARY OF EMP GENERATION. Gamma rays are scatteredfrom molecules with the emission of Compton electrons in theforward direction with energies on the order of 1/2 millionelectron volts. The motion of the Compton electrons ismodified by the geomagnetic field. They follow a spiralpath about the magnetic field lines until they are stoppedby collisions with atmospheric molecules. As the Comptonelectrons collide with atmospheric molecules, furtherionization occurs and the conductivity increases. Thepropagation of the EMP depends on the conductivity of theregion through which it passes. Dispersion, attenuation,and reflection may occur. The circular component of theCompton electrons represents a magnetic polarization, andthe linear component represents an electric polarization.Both the magnetic and electric polarizations vary withtime and, consequently, can radiate electromagnetic energy.This can also be viewed as a collective flow of electronsalong the field which radiates in the transversal direction.Compton electrons that move parallel to the geomagneticfield are not deflected. Thus, the EMP amplitude is smallin two directions along the geomagnetic field line passingthrough the burst point. The LMP amplitude is a maximum onthose rays from the burst point which run perpendicular tothe geomagnetic field in the source region. Since thelocation of enemy bursts cannot be predicted in relation toa system, one should harden for the worst case. Once again,the significance of the large geographical areas that highaltitude EMP can cover needs to be emphasized. The geo-graphical coverage as a function of the height of burst canbe obtained by considering the tangent radius. The tangentradius is the arc length between the line from the earth'scenter to the burst point and the line from the earth'scenter to the point where a line from the burst point istangent to the surface of the earth. See Figure 6. Fora burst at 300 kilometers, the tangent radius is 1920 kilo-meters, or about 1200 miles. The LMP from the burst cancover this region.

6. EMP EFFECTS. Most of the damage caused by the EIMPthreat occurs due to fairly simple effects. The strongelectromagnetic fields are converted into large voltagesand currents on any power lines, towers, cables, conductingloops, etc. These large currents and voltages, which rise

13

Mt

In-t

Ef£ @I

I.-

(wl) w

14

very rapidly, can destroy sensitive components and openprotective relays, thereby requiring repair, start-up, etc.Semi-conductor devices are particularly sensitive to voltageand current surges. The strong fields associated with EMPMlay destroy magnetic memory, cause logic circuits to berandomly disarranged (thereby destroying computer or otherelectronic device results at least temporarily) and causeall sorts of similar results. LMP also produces severedistortion and disturbance of radar and communicationssystems, at least during the pulse itself. The degree ofpermanent damage, if any, depends upon such factors asfrequency range, bandwidth, antenna size and orientation,and the types of components involved. Protection againstthis LMP threat thus involves several facets, each of whichis addressed further in this handbook. The effects of EMPmay range all the way from temporary interference, as witha coLuiunications channel or a computer in which a sequenceof calculations are caused to be in error, to permanentdamage, as would result if a semi-conductor or rectifier isburned out. Other intermediate effects such as computermemory destruction, requiring reprogramming and/or reini-tialization, or interruption, requiring system recycling tobecome operational, are also possible.

6.1 Llectric Field Effects. Examples of the electric fieldpulse generated by a high altitude burst will be given forthe case where the line from the center of the earth to theburst point makes a relatively small angle with the geo-magnetic field in the source region. This would be the casefor a burst at mid-to-high latitudes. The result is thatthe electric field at ground zero is relatively smallcompared to the field at the maximum field point (where aline from the burst point to the maximum field point isorthogonal to the geomagnetic field lines in the sourceregion).

The average distance in air at sea level that a gammaray travels before making a Compton collision is about 200maeters. A few gammas go as far as a few kilometers. Comp-ton electrons travel outwards only a meter or two beforebeing stopped by air. The radial current of Compton elec-trons is very intense near burst and decreases with radius.This current becomes negligible at a few kilometers. Thesource region for a surface burst is thus more dependenton the absorption of gammas in the atmosphere than on theyield of the weapon. Each of the Compton electrons inslowing down make many thousands of secondary electrons.

15

Thus, air becomes conducting. The outward displacementof electrons in air results in a radial electric field.-This radial electric field drives a conduction currentwhich tends to cancel the Compton current and therebylilits the electric field. The ground shorts out theradial electric field near it since the ground is normallya better conductor than air. Near the ground, conductionelectrons find an easier path by flowing down to the groundand back towards burst point. The result is a currentloop which generates an azimuthal magnetic field. A verti-cal electric field is required in connection with thevertical component of conduction current. This verticalelectric field can be regarded as connecting Compton elec-trons in the air with their image charges in the ground.Thus, for a surface burst, there is a radial and verticalelectric field and an azimuthal magnetic field. Irregular-ities in ground properties and surface features give othercomponents. The magnitude of the fields depend stronglyon distance from the burst and on the yield. Thus, EMPhardness specifications will depend strongly on the generallevel of attack to which it is desired that the systemsurvive. Hardness to other weapons effects, like blast,should be balanced with EMP hardness.

6.2 ElMP Effects Below Ground. The EMP fields below theground must also be considered. Electromagnetic fieldsdiffuse into the ground. The penetration depth dependson ground conductivity and frequency component considered.A greater ground conductivity leads to a smaller penetrationdepth. The high-frequency components penetrate less deeplythan the low-frequency components of the pulse. Thesource region where Compton current and air conductivityare important is a few kilometers in radius. Outside thisregion, the fields propagate like radio waves and fall offwith the reciprocal of distance. From a large distance,the source region looks like a conducting hemisphere thatis charged suddenly and then rings and radiates.

16

.1J

D. EMP PROTECTION

1. INTROWUCTION. For the design of a particular facili-ty, first the threat level, the survivability requirement,and the susceptibility level of the systems must be deter-mined. The operational and functional requirements of thefacility must also be considered; i.e., the design of EMPprotection for a facility which must continue to operateduring a nuclear detonation and for an extended periodpost-blast may be vastly different from that of a facilitythat needs only to be able to survive the nuclear blastand to operate for a short period post-blast. The level ofthreat and the frequency band over which the threat existsdepends on the weapon size, burst distance from the system,altitude, and several other factors. The values assumedfor these parameters will depend on the mission survivalrequirements established. A guideline for threat levelwould be the overpressure that the building can withstandand the EMP anticipated from aoburst yielding this over-pressure. To say a system is to survive the EMP for agiven threat level is definite only if the system and thethreat level are definite. EMP can exist independently ofblast if the burst occurs at a distance or at high altitudes,so protection may be required even for unhardened facilitiesmany Miles away from probable targets. The amount ofhardening required must be determined during design stagesbased on the above factors since the state-of-the-art issuch that "cookbook" approaches are not possible and theapproach used should be selected on the basis of eachindividual system. In the case of a high-altitude detona-tion, high energy waveforms, but somewhat less intense thanthose which appear in the ionized sphere, are radiated fromthe source region. Various responses of surface equipmentto this EMP can be observed, ranging from i"consequential"static" to burnout. The most severe effects are associatedwith the more susceptible components which are connected tolong exposed cables or antennas. One candidate for a severeeffect would be a transistorized shortwave receiver connectedto a large antenna.

2. EMP PROTECTION PHILOSOPHY (General). EMP protectionphilosophy is based on protection from three environmentalareas of concern: ground current effects, magnetic fieldeffects, and electric field effects. The solution to the

17

wrblem of ground current is to control the path of currentnxc sco I lected by water pipe, conduit, or other conductive

maaterials entering a facility or a structure, withoutcreating objectionable electric fields due to discontinu-ities or penetrations of such conductors. Ground currentsentering the vicinity of a structure through conductivematerials are conducted around the structure through aground counterpoise or structure shield and are dissipatedinto the earth or other conductive materials on the oppo-site side of the structure. If water pipes or conduitsare allowed to enter a structure without being connectedto the ground counterpoise, or other paths around the struc-ture, the ground currents would be conducted throughout thestructure, causing unwanted electromagnetic fields to bepresent in and near the structure. In order to prevent thepossibility of malfunction due to circulating currents inneutral wiring of transformers, such a system neutral, ifgrounded, shall be grounded at only one point IAW existingelectrical codes (i.e., National Electrical Code). Toprotect electrical wiring and components from magnetic andelectric fields, some type of shield must be utilized. Theshield must surround items to be protected. This type ofprotection does not completely isolate the item from theelectromagnetic field but attenuates the field strength toan acceptable level. Conductors may be shielded with someform of raceway, or conductor shields and cable armor.Components may be shielded with sheet steel housings.Ai entire structure housing electrical system may beshielded with sheet steel or nominal reinforcing bars whenproperly bonded and grounded. However, the use of rein-forcing bars for shielding is not very efficient and,except for cases where only a low level of attenuation isneeded, the amount of protection realized by this methodis generally insufficient and additional shielding will berequired. The cost to shield by reinforcing bars, whencompared to the cost of overall shielding for a given atten-uation, is generally more expensive and, therefore, unwar-ranted. Shielding against magnetic field pulses consists ofenclosing the region to be protected within an electricallyconductive shield. The magnetic pulse flux tends to con-centrate initially on the outside of the enclosing shell,progressively penetrating toward the inside of the shellas the pulsed field encompasses it. If the shell is elec-trically continuous, the voltage induced by the pulse fieldforces a current around the shell. The intensity of thevoltage and current surges produced by EMP must be reducedor attenuated to a level that will not damage or cause mal-function in the system being protected. This is accomplished

18

chiefly by reflecting the incident fields from the shieldedenvelope protecting the area, and by exponential absorptionof the residual currents induced in the protective shielding.These two basic methods of achieving attenuation may reducethe liMP to an acceptable level in some instances.

a. Reflection. The reinforcing bars, wire mesh and/or sheet metal housing utilized to protect the facilitysystems will reflect some of the incident EMP fields.Electric fields are reflected more than magnetic fieldsbecause of the greater difference in the wave impedancebetween electric fields and the inherently low intrin-sic impedance of metals. That which is not reflectedinduces a current in the shield and is exponentiallyabsorbed in the metal.

b. Absorption. Absorption is the principal methodused to attenuate EMP. As the impinging field is stronglymagnetic, thus a low impedance source, little energy isreflected from the shield. This causes large residualcurrents, eddy currents, to remain on the shield. The eddycurrents interaction with the shield absorbs some of thefield energy and dissipates it in the form of heat.

Consistent with past experience in radio frequency inter-ference and electromagnetic hazards to ordinance problems,the EMP hardness should be considered during preliminarysystem design and layout. lMP hardening requirementsshould then be kept clearly before the engineers duringadvanced design and system development stages. If the EMPproblem is not approached until late in system design, thecost and weight penalty can be enormous. One acceptedapproach is to incorporate as many of the simple liMP pro-tective measures as possible early in the design phase.The basis for this approach is that (with certain excep-tions) the costs for various protective techniques aregenerally small, provided that they are incorporated in anearly phase. Such low cost features include the use ofclipping circuits to protect sensitive transistors or theselection of noise-immune cable and grounding systems,which employ balanced twisted pair cables in shieldedconduit. If some shielding is required, little extra costis incurred to provide shielding against the most severethreat, since the electromagnetic performance requirementsis only one of the determining factors of how the system isdesigned and constructed. In most shielding situations,mechanical fabrication and corrosion protection costspredominat1. The shielding, therefore, can be conveniently

ovordesignod. Thus, it is considered more economical toinclude additional EI protective features rather than torisk rejection of the entire system during final EUP testacceptance procedures. This viewpoint is reinforcedbecause of the uncertainties involved with a number ofpractical considerations such as corrosion of joints,method of installation, and later modifications.

2.1 Application to Existing Facilities CGeneral).

a. Planning. Where costs are not prohibitive, itwill be desirable to apply hiM1' protective measures toexisting communications facilities, personnel shelters, orotherwise nuclear hardened buildings. Many present struc-tures were built before the existence of an DIP threat hadbn ralied. Ie the design of protective measures forsuch a large variety of structures, consideration of thefollowing items should be very important to the designerproviding EMP hardening for the structure.

(1) Threat Levels. The expected IMP1 throatlevels at the structure are based on type of burst, dis-tance to burst, and weapon yield. A guideline as to weaponyield and anticipated distance from a burst would be theoverpressures that the building has been designed to withstand.

(2) Susceptibility. A limited amount of suscep-tibility data is given infthis document. Existing EMPhandbooks and test reports are continuously being publishedproviding more information on susceptibility data.In general, only engineering estimates of system suscep-tibilities are known at this time. Some components ofsystems are known to be more susceptible than others basedon inherent voltage and current characteristics. Dis-cussion of general susceptibilities of some types of com-ponents which are common to a large number of systems aregiven. Power susceptibility of some components such asmotors, relays, transformers, or switches have beenmeasured.

(3) Cost of Retrofits. The level of funding avail-able as compared to the cost of the anticipated protectivemeasures may predominate in the determination of the extentof the lEMP retrofit program. In some cases, it may be moreeconomical to abandon a site and rebuild with plans whichinclude design of protective measures.

20

Ia

b. Exmination of Plans. A preliminary step in theprogram to EMP harden existing facilities is to examineavailable construction drawings and make a determinationof the construction methods used at the site. For example,if the construction is typically reinforced concrete, deter-mine the diameter, spacing or extent of use of rebar and,if any portion of the rebar was welded or brazed at jointsand intersections. A limited amount of shielding fromradiated fields will be provided by having bonded rein-forcing steel as part of the basic structure.

c. Site Survey. After an initial examination of con-struction awings and plans, it will be advisable to makea visual inspection of the facility; compare constructionto as-built drawings if available.

(1) Shielding. Any continuous steel plates willalso provide shieding and should be considered as part ofthe EMP protection inherent in the structure. The use ofadditional shielding such as form fitted copper sheeting,welded steel lining, or built-in place commercial shieldedenclosures should also be considered.

(2) Grounding. Determination of the availabilityand quality of thie grounding system for impulses and surgeswould be the next step in a retrofit program. Existence ofground connections to the basic structural steel and rein-forcement steel should be determined through resistancemeasurement. These grounds can be supplemented by addingan appropriately designed impulse ground counterpoisesystem to obtain surge grounds with resistances on the orderof about 10 ohms. Mleasurement of soil resistivity will aidin designing additional grounding.

(3) Commercial Utility Lines. It is usuallyanticipated that commercial utility lines will be destroyedand so a transfer switch will normally be available toswitch the load to the emergency power system. Thistransfer switch must be protected from damage from theinitial EMP encountered. A standard lightning arrester atthe transfer switch is normally sufficient for this purposeif connection to a low impulse impedance ground is availablenear the transfer switch and the switch is on the highvoltage side of a step down transformer. Typically, thenearest existing lightning arrester will be on the lastoverhead transmission line pole. The site survey shouldlocate and evaluate the quality of existing arresters,transfer switch grounds and distribution transformers.

21

(4) Ei.ergency Power. Vulnerable points of theemergency power system which require protective measuresshould be determined during the survey. These includeremote control conductors, remote indicator wiring, gener-ator exciter control, and battery charging circuits. Allsuch conductors should be contained in ferrous conduitwith threaded couplings. Power distribution from thegenerator must be similarly contained in ferrous conduits.Entries of this conduit into switches and junction boxesnear the generator should be suited for application ofradio frequency gaskets or electrically conductive bonds.

(5) Internal Power and Control Wiring. The meansof distribution of power conductors and control conductorswithin the facility should be determined by examination ofthe construction drawings and visual inspection of the instal-lation. The type of conduit used is important. Rigidferrous conduit with threaded couplings can provide shieldingof conductors while condulets give considerably less protec-tion and will require additional measures, such as theapplication of conductive filter-loaded plastic resins toeach joint connection. Numerous bends in conduit runs tendto degrade the shielding provided. The larger the conduitdiameter, the more effective it will be. Multiple runs ofconductors within cable trays tend to provide limitedshielding for each other. In the survey, such multiplecable runs should be examined for the possibility of instal-ling them within enclosed ferrous raceways. Entries ofconduit into switch boxes and junction boxes should beinspected for continuity, tightness, and quality of coup-lings. Box lids will require installation of radio frequencygaskets with appropriate treatment of the mating flangeswith conductive coatings. Open indicator panels and controlracks may require enclosure of wiring within an integralmetal cabinet, and the results of the survey should indi-cate the quality of existing panels, openings, and seals, aswell as existing potential for added shielding measures.

(6) Signal and Telephone Lines. Overvoltage pro-tection measures should be applied at the point where thelines first penetrate the shelter wall. All incoming lines(this includes power, telephone, and data lines) should bererouted so that at the entry point, protectors can bephysically mounted on a bulkhead which, in turn, can beelectrically connected directly as possible to an impulsegrounding system.

22

(7) Antenna Lead-in Cables. Antenna masts andcables should be examined for the existence of previouslyinstalled lightning protection measures.

3. ENERGY COLLECTION. The radiated electromagnetic fieldsfrom a high altitude nuclear detonation are important sincethese fields can appear over large regions. These fieldscan cause charges to flow on any good conductor. The way inwhich the energy is collected is often complex; but, ingeneral, the larger or more extensive the conductor, thegreater the amount of energy collected. For example, thewhip antenna of an automobile radio will collect far lessenergy than an AM broadcast transmitting antenna. Typicalcollectors of EMP energy include:

a. Long cable runs, piping, or conduit.

b. Large antennas, antenna feed cables, metallic guywires, or metallic antenna support towers.

c. Power or telephone lines.

d. Metallic structural building members such as girders,corrugated metal roofs, expanded metal lathe, or rebars.

e. Buried pipes or cables.

f. Long runs of electrical house or building wiring,conduit, etc.

g. Metallic fencing, railroad tracks.

4. RESPONSE OF COMPONENTS. The energy appearing in theelectromagnetic environment is converted, often in a complexfashion, into large amplitude currents and voltages flowingon any metallic conductor. To cause damage, it is neces-sary that these currents and voltages encounter a sensitivecomponent such as a transistor. In the case of an antenna,this would be the normal consequence of the function orpurpose of the antenna. In the case of other metallicstructures, various obscure details (such as quality ofwelds) control the distribution of currents and voltages.Electrical systems exposed to EMP may suffer degrada-tion in two distinct ways: (1) functional damage or (2)operational upset. If sufficiently large electric tran-sients are introduced, a component or a subsystem may becomepermanently inoperative until some part or parts are replaced.If a system is permanently damaged in this manner, it is

23

said to have suffered functional damage. Other types offunctional damage may occur wherein a particular deviceor subsystem is rendered only partially capable of executingits entire range of functions. Another aspect of functionaldamage is the decrease in the lifetime of a particularcomponent or subsystem. Electrical transients may tem-porarily impair the performance of a system. This impair-mlent may last for only a few microseconds or could be hours.This temporary impairment of the system's operation is knownas operational upset. The importance of either functionaldamage or operational upset within the system depends uponthe specific characteristics of the system. Beginning withthese definitions of degradation, it is useful to considerexamples of each type of effect. Burnout of a transistoror the opening of a fuse are clearly two examples of functionaldamage. Examples of operational upset are the erasures ofmagnetic core memories of computer systems or the resettingof flip-flop circuits. Depending on system design, theunanticipated change of circuit conditions or temporarymalfunctioning of a number of control devices could rangefromi insignificant to catastrophic. Electronic componentsare often very sensitive to functional damage or burnout.These are listed in the order of decreasing sensitivity todamage effects:

a. Magnetic core memories (erasures).

b. Microwave semi-conductor diodes.

c. Field-effect transistors.

d. Radio-frequency transistors.

e. Audio transistors.

f. Silicon-controlled rectifiers.

g. Power rectifier semi-conductor diodes.

h. Vacuum tubes.

Thus, systems employing vacuum tubes are far less suscep-tible to EMP effect than those employing transistors.Various electronic or electrical systems are subject tomalfunction:

24

u. Most Susceptible:

(1) Luw power, higIh-sl)ecd digital computer (upset)cither transistorized or valcuum tube.

(2) Systems employing transistors or semi-conductorrectifiers (either silicon or selenium), such as:

(a) Computers.

(b) Computer power supplies.

(c) Transistorized power supplies.

(d) Semi-conductor components terminatinglong cable runs, especially between sites.

(e) Alarm systems.

(f) Intercom systems.

(g) Life-support system controls.

(h) Some telephone equipment which is partiallytransistorized.

(i) Transistorized receivers.

(j) Transistorized transmitters.

(k) Transistorized 60 to 400 cps converters.

(1) Transistorized process control systems.

(m) Power system controls.

(n) Communication links.

b. Less Susceptible:

(1) All vacuum tube equipment (does not includeequipment with semi-conductor or selenium rectifiers):

(a) Transmitters.

(b) Receivers.

(c) Alarm systems.

25I.

(d) Intercoms.

(e) Teletype-telephones.

(f) Power supplies.

(2) Equipment employing low current switches,relays, or meters:

(a) Alarms.

(b) Life support systems.

(c) Power system control panels.

(d) Panel indicators, status boards.

(e) Process controls.

(3) Hazardous equipment containing:

(a) Detonators.

(b) Squibs.

(c) Pyrotechnical devices.

(d) Explosive mixtures.

(e) Rocket fuels.

(4) Other long power cable runs employing dielectric

insulation, equipment associated with high energy storage

capacitors or inductors.

c. Least Susceptible:

(1) Hfigh voltage 60 Hz equipment:

(a) Transformers.

(b) Motors.

(c) Lamps.

(d) Filament.

(e) Heaters.

26

(f) Rotary converters.

(g) Heavy duty relays.

(h) Circuit breakers.

(i) Air insulated power cable runs.

The less susceptible equipment or components would be rtiademore susceptible if they were connected to long exposedcable runs, such as intersite wiring or overhead exposedpower or telephone cables.

5. DESIGN PRACTICES. In previous sections, a discussionof EMP generation and effects, general protection philosophyand general response of components was presented which wasintended to give the reader a basic understanding of EIPand the magnitude of providing adequate MP protection.The rest of this handbook will concentrate on providinga basic design practice for LMP. The information is basicin nature and only highlights major areas that need to beconsidered in basic lIP protection. Much of the informationin this section was obtained from a DNA Awareness Courseon UIP and provides a good starting point for consideringLMP protection. This section will emphasize systemsengineering from the start and considers the entire systemdesign first and then breaks up into individual componentdesign.

5.1 Systems Aspects. Perhaps the most disconcerting fea-ture of most system EMP programs is the magnitude of ahardware assessment. Some loose physical criteria need tobe applied early, in order to guide the sorting and thechoosing. As is often the case in such situations, thefirst steps may be partly artificial. To start with,the zoning approach in structuring a system and its EMPanalysis will be considered. Before starting though, varioussystem definitions should be introduced.

5.1.1 System Level Concepts and Definitions. It is impor-tant to keep in mind that when we say "system", this canencompass a broad spectrum of configurations in size, shape,and complexity. A pocket transceiver is as much a "system"as is an ABM radar site. Thus, its definition is: acomplete, self-contained primary-mission entity. Someother system-level concepts and definitions are noted here:

27

I i .l l I i l I I I "II

it. Zoning: The identification and integrations ofregions of similar electromagnetic (EM) environment and/orsusceptibility.

b. Clustering: The grouping of elements of similarcharacteristics and purposes.

c. Layering: The sequencing of zones and protectivemeasures between outer environment and inner equipment.

d. damping: The use of lossy elements or materials toabsorb EM energy.

e. Violations: The features which represent defectsfrom a systems hardness viewpoint.

f. Fixes: Obviously, the measures taken to rectifyviolations.

5.1.1.1 Zoning. Considerations of EM zones within asyste nearly always appear overtly in terms of shieldingeffectiveness. Zone boundaries are generally constrainedto coincide with major geometric or structural contours, orwith intentionally introduced shields or shielding enclosures.Thus, it is usually the case that zoning considerations donot appear explicitly in an EMP systems analysis. However,there have been system cases in which the EM geometrieswere so intricate that elementary shielding considerationswere obviously inadequate. It was then essential to performa meticulous mapping of the EM zones. This has been particu-larly true in certain nuclear test situations.

a. Zoning Approaches. Electromagnetic (EH) zones maybe defined in two broad ways:

(1) Environmental zoning, in which the magnitudeand shape of the field pulse is defined within the succes-sive regions from outside in.

(2) Susceptibility zoning, in which the magni-tude and frequency (or timeT domains corresponding to thevulnerability thresholds are scaled from inside out.

In a "good" system (See Figure 7), zone boundaries appearas (nore-or-less) concentric iso-contours and there is an [appropriate coincidence between environmental and susceptibilityzones.

i28

28

.0-

IrwD

U -CD

42

b. Zone iDefinition. It is conuaon practice to definethe levels of different zones in terms of relative dB.Physical features (such as walls, cabinets, etc) with whichzone boundaries are iost usually associated, generallyappear to be good for 20-30 dB, without specific EMP-orientedtreatment. Zones can also be delineated in terms of otherelectromagnetic interaction specifications. For example,equipment tested to meet FED STD 461/462 or FED STI) 222-Abelong in one of the better shielded zones.

5.1.1.2 Clustering. Evidently, one of the things whichshould improve EMP hardness is the reduction of the areaover which vulnerable elements are located. All othersystems aspects being equal, it is generally best to con-tour the EMP zones as compactly as possible. This isespecially important if upset, such as computer memoryerasure, is concerned.

5.1.1.3 Layering. Most of our simple examples here showEMP protection as appearing in several successive geometricstages or layers. Of course, each boundary has to be com-plete in the sense that apertures and penetrations must betreated to preserve what was gained at that layer (remem-ber the weakest link). There seems to be a tendency todeal with EMP at one (or at most, two) boundaries. In somesense, EMP tends to be seen like "plant security". Put upone good, well-patrolled fence. But, in many EMP cases,this is quite unrealistic. For instance, in a deeply buriedsystem, it is plainly obvious that some protection can begained almost "free" from the earth cover itself. It isalso unrealistic in "porous" systems -- that is, systemswith very many apertures and penetrations.

5.1.1.4 Ringing. There are two wholesale approachesto Ell field protection. One of these is the "iron curtain"method, in which the various elements are thoroughlyshielded and electrically isolated from one another. Theother is the "common sink" in which the various elementsare massively connected together. The difficulty is thatone really cannot do either thing thoroughly. Elements mustbe connected together somehow, but they cannot all be placedin intimate contact. So, the result is something in between,which often ends up acting like a high-Q EM cavity, or LCcircuit. This is basically why many partially shieldedsystems exhibit strong ringing when excited by means of anEMP stimulator.

30

5.l.l.S Damping. Such ringing represents efficient storageof EM! energy and a prolongation of the time during which itcan be coupled to internal elements and circuits. We canreduce this condition by spoiling the Q of the enclosuresand implicit circuits. The concept is primitively illustratedhere in the insertion of (parallel) damping resistors.(Series damping requires careful circuit analysis to avoidmiaking matters worse.) This technique has been very success-ful in some types of nuclear test interference problems.Most shielding systems have characteristic impedances in therange of 5 to 200 ohms. Damping resistors of correspondingvalue are used; the exact choice is not critical. Thistechniquv is mlost appropriate wherein the shielded enclosureis (for a variety of reasons) poor, thus permitting entryof the higher frequency (ringing-frequency) components.

S.1.1.6 Violations and Fixes. The zoning concept hasanother advantage in complicated system evaluations. Itpernits the definition of specific locations and components(along a boundary) requiring liMP treatment. In the strictestanalytic sense, one assigns a minimum dB tmargin which allpoints and elements within and at such a boundary are tosatisfy. Those that don't are at once identified as "violations".its we said before, "fixes" clearly encompass those measurestaken to redress these situations, or in some cases, toredress their consequences. Violations generally fall inone of four broad classes as outlined here. The subsequenthardware categories are addressed to rectify one or anotherof these general conditions:

a. Circuit Considerations.

b. Shielding.

c. Cooling.

d. Grounds.

e. Protection and Testing.

S.l.l.0.l Circuit Considerations. Extensive studies havebeen made on evaluating the effect of ElMP energy in elec-tronic circuits and components. Chapter 5 of reference 27,EIiIU Handbook for Missiles and Aircraft, provides additionalinform ation on component vulnerability and should be con-sulted for further information. One of the main concerns,however, is how I3MP energy gets into circuits. One method

31--

. . ....

is coupling. Coupling should be avoided in circuit layouts- most notably, inductive loops. Good circuit layout prac-tices apply to large cable systems or to printed circuitpackage. Another is by injection as extraneous voltage orcurrent pulses at peripheral terminals. To solve theseproblems involves detailed circuit analysis techniques andspecial computer programs, and it is felt that furtherdiscussion in this area is beyond the scope of this basichandbook. References 22 and 27 should be consulted for furtherinformation.

5.2 Shielding. For comparative purposes, it has becomecustomary to rate a design or product in terms of "shieldingeffectiveness", measured in dB. This is a somewhat ambigu-bus term since it will depend, in practice, on the sizeof the box, the location of the item, the frequency domainand the method of measurement. Basically, it can be viewedas a measure of a certain internal environment "with"versus "without" the protective scheme.

5.2.1 flow Do EMP Shields Work? In the EMP frequency domain,a douinant mechanism in shielding effectiveness is insidecancellation or field reflection due to induced surfacecurrents as illustrated here. Thin walls, higY resistancepaths, apertures, seams, etc., seriously affect the reflec-tion or cancellation characteristics and serve as internalfield generators as well. See Figure 8 for typical shieldingdesign problems. A good shield must, therefore, be sufficientlythick, continuous, complete, and tight. This is essentialfor shielding above 60 dB.

5.2.2 Ideal Versus Realistic Shielding. Of course, mechan-ical and electrical inputs and outputs are also essential,and economic realities place real limits on wall thickness.Shielding hardware considerations generally boil down tocompromises in relation to:

a. Wall thickness and material.

b. Apertures -- tightness.

c. Penetrations -- conductors.

5.2.3 Shielding Effectiveness. The plane wave theory(or transmission line theory) of shielding is the basis ofthe most commonly used shielding design data. The resultingset of design equations is based upon the separation of theshielding effectiveness into three additive terms: absorption

32

-~, 'V I II I II II

......

4 Iif

33

loss, reflection loss, and a correction term to account forre-reflections within the shield.

The shielding effectiveness (in decibels) of a large, planesheet of metal with an EM wave arriving along a path perpen-dicular to the sheet has been shown to be:

SE - 20 logleYl + 20 log.1J + 20 logll - re2YLI,

A R C

where: L - thickness of the shield.y - propagation constant of the shield.- transmission coefficient, and

r - reflection coefficient.

The shielding equation is often written as:

SE - A + R * C [2]

where A, R, and C are the indicated three terms in Equation1 and represent, respectively, the Absorption Loss, the

Reflection Loss, and the Correction Tirm for re-reflectionsas discussed earlier. In-a particular shielding application,the values of the constants y, r, and T depend upon theconductivity (a), permeability (p), and permittivity (c)of the shielding material. The values of r and T dependalso upon the wave impedance of the EM wave impinging uponthe shield. For convenience in the use of the shieldingeffectiveness equation, the individual terms A, R, and Chave been expressed in more readily usable forms as func-tions of the EM wave's frequency (f) and of the shield'sthickness (), relative permeability (p ), and conductivityrelative to copper (g Simplified approximate expressionshave been derived for the reflection and correction terms.The selection of the appropriate approximate expressionwill depend upon whether the wave impedance is low (Z <<377;magnetic field), medium Zw 377n; plane wave), or higW(Z >>377i; electric field). Low impedance fields are found

inwthe proximity of loop antennas, high impedance fields arefound near dipole antennas, and plane waves exist away fromthe near fields of source antennas.

5.2.3.1 Absorption Loss. The absorption loss of an EM wavepassing through a shield of thickness 4 can be shown to begiven by:

A KCVP ()3)

34

where: K1 - 131.4 if t is expressed in meters, or3.34 if t is expressed in inches.

f - wave frequency, liz.I - shield thickness

Or - relative permeability of shield material, andgr - conductivity of shield material relative to

copper

Note that the absorption loss (in decibels) is proportionalto the thickness of the shield and also that it increaseswith the square root of the frequency of the EM wave to beshielded against. As to the selection of the shieldingmaterial, the absorption loss is seen to increase with thesquare root of the product of the relative permeability andconductivity (relative to copper) of the shield material.

Table 1 contains a tabulation of electrical properties ofshielding materials (gr and Or); since Or is frequencydependent for magnetic materials, it is given for a typicalshielding frequency of 150 kHz. The last two columns ofTable 1 evaluate Equation 1 to give the absorption loss at150 kHz for both a one millimeter and a one mil (.001 inch)thick sheet for each of the listed materials. The absorp-tion loss for other thicknesses can be calculated by simplymultiplying by the shield thickness in millimeters or mils.Shield thicknesses are commonly expressed in either milli-meters (mm) or milli-inches (mils); these two units arerelated as follows:

1 mm - 39.37 mils or 1 mil - .0254 mm.

The variation of absorption loss with frequency, as well asa comparison of the absorption loss of three common shieldingmaterials one mm thick, can be seen in Table 2. Alsoincluded is a listing of the relative permeability, as afunction of frequency, for iron.

Remember that the absorption loss is just one of three addi-tive terms which combine to give the attenuation (shieldingefficiency) of the shield. At this point, the absorptionloss has been presented in equation form (Equation 3) andtabular form (Tables 1 and 2). The tabular forms are easy-to-use sources of accurate results when the shield material andfrequency of interest are included in those tables and graphs.Quick results for almost any material and frequency combination canbe obtained from an absorption nomograph,'but the results

35

are generally less precise; nomographs are a good sourceof data for initial design purposes. Once a shieldingmaaterial and thickness are tentatively selected, one maywish to compute a more precise value of the absorption lossby evaluation of Equation 3.

5.2.3.2 Reflection Loss. According to Equation 1, thereflection loss portion, R, of the shielding effectiveness,SE, is given by:

R - -20 log HdB 4

Where: t is the transmission coefficient for the shield.The reflection loss includes the reflections at both surfacesof the shield and is dependent upon thewave impedance and frequency of the impinging EM wave aswell as upon the electrical parameters of the shieldingmaterial. It is independent of the thickness of the shield.

In a manner analogous to the classical equations describingreflections in transmission lines, the shield reflectionloss can be expressed as:

R- 20 log 1 + S dB[14SI dB[]

where: S is defined as the ratio of the wave impedance tothe shield's intrinsic impedance and is analogous to the vol-tage standing wave ratio in transmission line practice.While the shield's intrinsic impedance is easily deter-rined from the electrical properties of the shield material,the wave impedance is highly dependent upon the type andlocation of the EM wave source.

5.2.4 A Shield is a Magnetic Field Reducer. Let usneglect apertures and penetrations, so that the internalfield inside a shield is determined by overall field pene-tration or "diffusion". Assume further that the incidentfield is more-or-less "white" in frequency content, as forEMP. Then the internal waveform will tend to be dominatedby frequencies just below the attenuation cutoff, and, ofcourse, this effect would be accentuated in internal mag-netic field.

5.2.5 Diffusion Shielding. There is another kind ofshield -- the semi-permeable type. Examples: earth cover,rebar grids. Here the effective skin depths may be large,and the total attenuation relatively small -- like 30 dB.

36

'IR '

TABLE 1

ELECTRICAL PROPERTIES OF SHIELDING MATERIALS AT 150 KIIZ

Relative RelativeConductivity Permeability

Metal gr Pr Absorption Loss

(dB)

1 nmm thick 1 mil thick

Silver 1.05 1 51.96 1.32Copper, annealed 1.00 1 50.91 1.29Copper, hard-drawn 0.97 1 49.61 1.26Gold 0.70 1 42.52 1.08Aluminum 0.61 1 39.76 1.01Magnesium 0.38 1 31.10 .79Zinc 0.29 1 27.56 .70Brass 0.26 1 25.98 .66Cadmium 0.23 1 24 Al .62Nickel 0.20 1 22.83 58Phosphor-bronze 0.18 1 21.65 .55Iron 0.17 1,000 66S.40 16.90Tin 0.15 1 19 69 .50Steel, SAE 1045 0.10 1,000 509.10 12.90Beryllium 0.10 1 16.14 .41iypernick 0.06 80,000 484.00* 88.50*Monel 0.04 1 10.24 .26Mu-metal 0.03 80,000 2488.00* 63.20*Permalloy 0.03 80,000 2488.00* 63.20*Steel, stainless 0.02 1,000 244.40 5.70

*With no saturation by incident field.

~ 4*~37

TABLE 2

ABSORPTION LOSS, A, OF I MM METAL SHEET

Frequency Iron Copper Aluminum

r A A A

r- r rB) TUrT60.0 H1z 1,000 13 1 1 1 0.8

1.0 kliz 1,000 54 1 4 1 3.0

10.0 kliz 1,000 171 1 13 1 10.0

150.0 kliz 1,000 663 1 56 1 40.0

1.0 MHz 700 1,430 1 131 1 103.0

3.0 M11z 600 2,300 1 228 1 178.0

10.0 Miz S0 3,830 1 416 1 325.0

15.0 M1hz 400 4,200 1 509 1 397.0

100.0 M11z 100 5,420 1 1,310 1 1,030.0

1.0 G1Iz 50 12,110 1 4,160 1 3,250.0

1.5 GtIz 10 6,640 1 5,090 1 3,970.0

10.0 GiiZ 1 5,420 1 13,140 1 10,300.0

Relative Conductivity, gr: Iron - 0.17, Copper- 1.0,Aluminum - 0.61.

38 iI l

Usually, this is used in combination with smaller, internal,and more complete shields. Often, such a shield appearsas a zone enclosure of opportunity, such as in buried orheavily reinforced structures. The waveform appearinginside such a diffusion zone will generally be a combinationof a short spike (possibly associated with apertures) and alonger "tail", related (as before) to the induced skincurrents in the conductor.

5.2.6 Apertures. Of course, there are many different kindsof "apertures"; but most importantly, they may be dividedas intentional and as unintentional. Perhaps the singleworst class of violation of good EMP protection practicearises in the accidental or unintentional compromise ofshielding integrity. Anything which interrupts the skincurrent path on a shield increases its impedance and actsas well as a radiator into the internal region. Hence,the effect of a seam crack is not measurable simply by itsphysical area, which may be quite small. If it is near aregion of high surface current concentration, it can coupleenergy to the interior many times greater than one mightexpect. In particular, physical breaks - such as seams andbonds, however well made - represent a constant threat tointegrity and protection value.

5.2.6.1 Seams. Of course, it is almost impossible tofabricate a shield as a single, unbroken, electromagneticenclosure. Large system enclosures can only be constructedby assembling large numbers of sheets or plates. Technically,the contact lines or seam between such single piecesrepresent potential apertures. The most common large-scaleseam techniques involve welding for steel and soldering orbrazing for copper. These fabrication methods in themselvesplace certain minimum thickness criteria on the material.Thinner sheets would "burn through" too easily. So, we seethat at least two mechanical aspects already impose minimumthickness requirements, which may, in fact, be greater thanrequired by just the EMP criterion by itself; strengthand fabricability. Such thicknesses run from 60 to 300 milsfor medium-large military construction. Overlap should pref-erably be 10-20 times thicness for thin sheets. Buttjoints can be acceptably used for thick plate, but thisusually requires welds on both sides, with careful probe testsfor weaknesses.

39

5.2.b.1.1 Dilemmas. A system designer may easily be facedwith a dilemma. Put in a shield - and it is likely to bemuch better than is really necessaryl But this held stillanother conflict - the necessary mechanical thickness providesan implicit (and high) dB protection value. Inexpensiveassembly methods, such as tack welding, seriously erodethat member due to aperture leakage at the long open seamsthroughout the structure -- it is then not much betterthan welded rebar. The "good shielding" criterion thusturns out to require continuous and meticulous weldingalong all seams in order to match the protection valueinherent in the material itself.

5.2.7 Gaskets and Bonds. Considering the difficultiesencountered with such seemingly "tight" apertures as weldedseams, it is no surprise that netal-to-metal contact sur-faces, held together by simple mechanical pressure, canconstitute serious violations of shielding integrity. Suchcontact areas are unavoidable at functional apertures(e.g., access doors, service hatches, equipment panels, etc).The terminology of "bonds" and "bonding" suffers from indis-criminate definition. It is used for two hardware topics:

a. Treatment of contact surfaces at extended electro-mechanical junctions (seam bonds).

b. Low-impedance interconnection of shields andcomru on reference surfaces (bond straps).

5.2.7.1 Seam Bonds. There is extensive literature on allmanners of bonding long, continuous, metallic, contactlines. They deal with a range of bonding permanency, frompermanent once-made joints through rarely-disturbed servicepanels, to continuously-exercised doorways. Of course, thelatter represents the most difficult problem in dependabilityand maintainability. The basic mechanical requirements forsimple reliable seam bonds is absolute flatness and elec-trical cleanliness. Neither of these is generally achievablein other than ideal laboratory conditions. The pragnatichardware problem is then to obtain low-impedance continuouscontacts at an acceptable level of "dirtyness" and "deformation".

5.2.7.2 Clean Contacts. Electrically clean surfacescan be readily obtained with pure tin, gold, palladium,platinum and silver. But zinc, plain cadmium, and very thingold platings are considered as acceptable substitutes.Easily oxidized materials (like Al) should be avoided.Lubricants are capricious. In some cases, they will inhibit

40

corrosion and oxidation and facilitate good metal-metalcontact. More likely (especially motor oils), they willdo just the opposite. Any plating is better than none,and controlled roughness is generally better than smoothsurfaced (machine scoring and knurling).

5.2.7.3 Pressure Contacts. Of course, roughness is oneway to compensate for surface irregularity. An ultimateway to do this is to use deformable conductive gaskets.A good way to understand the pressure contact problem isto consider a panel seam, bonded by means of bolts andflange strips. In the frequency domain of interest here,seams of this type require specific contact pressures of60 to 100 pounds per lineal inch for 80-100 dB attenuation.Obviously, this form of seam is best for "once-only" cases,which are expected to be broken very rarely, if ever, duringthe system's life.

5.2.7.4 Electromagnetic Gaskets and Panels. People alsoresorted to the "gasket" solution for "bonding" peripheralcontacts which would only be occasionally broken. It isalso useful for irregular or deformable surfaces. Thereare two "fairly" good types:

a. The flat molded metal gasket which deforms slightlyunder pressure. This is "throw-away" in the same senseas an engine head gasket.

b. The braided cord gasket. A variety of exoticdesigns appear on the market. The good ones from an atten-uation standpoint utilize deformable metal cores. Unfor-tunately, these have low resiliency and, at best, can onlybe reused two or three times. The synthetic core, doublebraid gaskets are generally more transparent in a givengeometry. The single braid types can only be describedas "abominable". Braided gaskets are not recommended forexposed, unmaintained situations. It is a lot of workto remove the crud which builds up around them and eventhen, you will probably be left with oxidized and corrodedspots mll along the contact line.

5.2.8 Shielded Enclosures. Structural flexibility withreliable shielding effectiveness has been designed intocommercially available shielded enclosures. These shieldedenclosures are being used for RF testing and measurement.At present, at least, it is not likely that military systems

41I

would employ such components except as accessories inproduction and testing phases. The prefabricated bolt-together enclosure has enjoyed wide acceptance. However,it does require periodic maintenance. The frame shiftscause open slits and metal-to-metal seam corrosion. Wherehigh shielding requirements exist, serious considerationshould be given to the welded seam enclosure.

5.2.8.1 Finger Stock and Doors. Resilient finger stockis a favorite solution for doors and hatches which mustbe frequently used. Here we see that it should be usedin double rows. Some writers suggest that the rows shouldbe staggered for maximum attenuation so that the fingers inone are opposite the slots in the other. At the higherfrequencies, this seems reasonable when one considers theradiation pattern of each slot seen as a tiny dipole.Finger stock is probably the most difficult protectionhardware to maintain. Traffic inevitably brings with itdirt and abrasion. The doors and frames must be extra stiffif the fingers and the contact surfaces are to maintaintheir register.

5.2.8.2 Protection Maintenance. This is a good placeto emphasize the importance of adequate EMP protectionmaintenance. There is probably no hardware as suscep-tible to wear and tear as aperture components. Some ofthe ways in which this can deteriorate and degrade theprotection factors are shown here. Formal procedures shouldbe adopted and adhered to in opening and resealing hatchesand access panels -- also for periodic service to personneland materials entrances.

5.2.9 Open Apertures. Apertures which could be "closed"electromagnetically by means of conductive materials (sheets)and construction similar to the surrounding shield havebeen considered. The significant problem was the peripheralcontrol. But some mechanical requirements call for aphysically open aperture for such things as ventilation,microwave lines, etc. Two broad classes of "solutions" arecommon for these - screens of various types and "waveguides-beyond-cutoff". The latter can also be used sometimes forentrance passages and doorways to avoid the finger stockproblem, where penetration of high frequency content isclearly not a problem.

42• , 7.

$.2.10 Screens. Ordinary heavy duty screening is goodfor the order of 40 dB. The trouble with ordinary mate-rials lies in corrosion and oxidation which can breakthe contact between individual wires. Electromagneti-cally, an old piece of screening may be a good coupler.The specially fabricated materials like "electromesh" aretreated to resist this action. Hexcel is usually okay, butthere have been instances of poor quality control in whichthe glue between the foils acted as an insulator. True"honeycomb" screening provides the best compromise betweenshielding and air flow. Where that is important (note cadplated AE/steel type), best results are obtained with honey-comb which has soldered, brazed or welded contacts betweenfoils.

5.2.11 Waveguide Schemes. The "waveguide-beyond-cutoff"is somewhat of a misnomer. Over most of the IMP frequencydomain, such a geometry is really behaving more like aquasi-static "field-bender". Indeed, it works even betterif we can put a 900 bend in it. The idea is to design it sothat its cutoff is significantly well above the high-frequency "roll off" in the environmental spectrum. This isnot difficult to do if it is under many feet of earth or ifit is already protected by some partial attenuation, such asa welded rebar cage. These situations tend to move the roll

off to lower frequencies, as we have indicated before.The approach is fine for ventilation, but don't make the apertureinto a propagating structure by running cables through it.

5.3 Penetrations. Again, there are many kinds of "pene-trations". Most commonly, one thinks of an insulatedconductor passing through an aperture. It may be carryingpower or functional signals, but uninsulated, "grounded"conductors, such as motor shafts, can also representpenetrations. Thus, there are two broad classes --electrical and mechanical. We see here ways in whicheach can provide paths for coupling and transferring energyfrom the external to the internal zone. Note particularlythat mechanical penetrations can be deceptively protected byinnocent-looking bonds, which are really high-impedancecouplers.

5.3.1 Electrical Penetrations. The existence of a true"electrical penetration" corresponds to an intentional orunintentional violation of the zoning concept. If anelectrical circuit is carefully confined to a single EMzone, then its penetration through a shield does not, infact, constitute a violation. In principle, it cannot

43

transfer energy which would not be there in its absence. Wemake this seemingly simple point to emphasize the necessityfor observing zonal hierarchies in proving conductor andcable shielding (as discussed in a succeeding section).But what about unavoidable conductor penetrations such as,for instance, long wire antennas? One thing to do torectify such situations is to provide entrance protectionin the form of filters or active devices (zener diodes,spark gaps, etc). These are discussed in the section on"Protective Devices". These protective devices should belocated in vaults or small shielded boxes. Finally, if allelse fails, one can isolate that portion of the system(which really goes back to systems and circuit layouts) andsimply make its terminal circuits very hard in themselves.So we see that the treatment of purposeful electricalpenetrations is not really a "shielding" topic. Table 3provides a summary of penetration treatments for communi-cation facilities.

5.3.2 Mechanical Penetrations. A conductive metallicpenetration may be deceptively protected. Consider, forinstance, a shaft passing through a bushing. Here we seetwo equivalent analyses for the problem of energy transferthrough such a seemingly "tight" geometry. In the highfrequency domain, it can be treated in terms of a circuitequivalent in which the important features are the bushingcontact resistance and bond inductance shunting to thecommon, shielded reference level. By these avenues, it isnot difficult to get a 30 dB leak in a 60 dB shield.

54 Cables. This subject tends to bring about thereaction: 'Cables are cables -- what can you do aboutthem?" Generally speaking, there is not much that canbe done once you have bought them; protection is bestincorporated in specification and in installation. GoodEMP practice probably costs relatively more in cables thanin just about any other hardware area. But without it,"cables" usually turn out to be a lingering sore spot.

5.4.1 What is a Cable? Basically, it is a collectionof insulated wires, each of which provides a low resis-tance (or low impedance) conductive path between twoseparated terminals. This is also sometimes labeled a"connection".

5.4.1.1 Cable System Design. The development of EMPcriteria for cable specification does not necessarilystart with the de facto system connection diagrams. Rather,

44"

c~ E

060

0. 004 , . 4

- 4-. VA"~

c in00j 0 4 0 ~ IV6~~" tv04, C 4

tU.-

<- C 0 0A 00j 4, C2

0 ma 0 0 0

4- 4, :2 4, 4,

0 o E 00-o 4- 0 o 0o 0-

ca 00

w C "0w L. 4.

4,

IA 00

0- 4, VIA m-

4, -4 4,4 c c %-08 4- 4) 04,4

c 0 0 0 .

o -A

0 0E.. -UJ &. 8 00 E o00

o 00V~~ 2 0

~~C 4-- 4

~~I 4. 0EC.C 4u. lb or- 4- "IOh L.0 0 Oh E 0 ~ 0 ~ 45

it should start in the determination of what is to beco-inected to what, and in how this is to be accomplished.As we will shortly indicate, much may be done to ease thecable hardening problem by more judicious circuit manage-ment. Of course, once the inescapable connection require-ments are determined, then one must get to the specifichardware issues.

5.4.1.2 System Configuration Options. One broad cate-gory of "fixes" involves reducing the severity of criteriaimposed on system cables. We can do this iy keeping inmaind certain tests for each cable and each wire in it:

a. Do we really need the functions served by thewire (or cable)?

b. Can we accomplish it by a "non-wire" technique?

c. All other things being equal, what options areopen for satisfying the connection requirement?

5.4.1.3 High Operating Levels. One way to "ease the strain"in cable protection costs is to work with high power andsignal levels in the longer cable runs. If this is notcompletely possible, it may be more economical to run thelower level circuits in separate and smaller, better-protectedcables. As suggested here, do not work with millivoltservos if it is just as easy to work in the volt range.

5.4.1.3.1 Component Distribution. Closely related tooperating level is the distribution of components. Asimple example appears here: The preference is obvious,all else being equal. Note, too, that the choice reflectson the character of the terminal circuits as well. Inthe preferred configuration, the pre-amp outputs andmonitor inputs will tend to be "harder" simply becausethey must operate at higher signal levels in their ownoperation.

5.4.1.4 Carrier Systems. Another way is to go to carriertechniques. Many sensing and control situations lend them-selves to this by relatively inexpensive terminal hardware- provided this choice is made soon enough. There existssome older system examples in which the additional criterionof IMP hardening would have easily tipped the scales infavor of carrier systems, rather than DC or low level,self-generating circuits. Carrier systems have the advantages

46

of permitting floating, balanced conductors, narrow bandpassfiltering, transformer isolation, easier nullification, andmuch more.

5.4.1.5 Circuit Arrangement. Generally, the EMP suscep-tibility of a cable system is related more to the sensitivi-ties of the terminal elements and circuits than to "breakdown" or "burnout" limits in the cable itself. Of course,this points at once towards terminal protection but a circu.tdesigner may be able to improve matters by keeping EMP inmind when he considers the terminal element design and thecircuit routing through the cable system.

5.4.1.6 The Zoning Role. It was mentioned earlier that dif-ferent zones may remain isolated, even though connected bylong cables. It is worth briefly pointing out why. First,the external propagation impedance (for short pulses) along along cable is such that it does not significantly alter theexternal zone coupling from what it would be in its absence.One need only be concerned with VLF (very low frequency,3-30K Hz) common modes. Second, the propagation attenuationdue to radiation or earth damping (if buried) is such thatanything originating at A is lost by the time it gets to Bin the usual case. Of course, if the cable is involved in a"large loop" geometry, more careful examination is in order.The above "solutions" are somewaht "sophisticated" and arerecommended only for the most carefully thought out and con-trolled situations.

5.4.2 Cable Fabrication. There are all sorts of ways ofmaking wires, insulating them, shielding them, and bundlingthem up. Some may be excellent for their primary purpose,but not for EMP.

5.4.2.1 Cable Types. When one examines the possible permu-tations of cable component choices, it is obvious that nodetailed case-by-case evaluation is possible. Rather, certainpreferable choices in each category of component will be dis-cussed. These can be broken up into two broad categories:Those aspects which influence the control of the effect ofexternal environment, and those which control the inner cableenvironment, circuit intercoupling, and so forth.

5.4.2.2 Why is the "Exterior" So Important? The dominant mecha-nism in shielding is the induced surface current. This is concen-trated in a surface layer measured by a "skin depth". Skin depth

L1.4

is the thickness of metal in which the effective "chargepacket" is traveling. For typical EMP-like pulses, this maybe 5 to 50 mils in thickness. Of course, if we want 60 to80 dB of protection, the metal has to be several times thisthick. As we see here, if it is too thin, the pulsediffuses right in and develops a field to the inner con-ductors. Of course, if there is no outer shield, this cur-rent pulse would be concentrated directly on the outerconductors of a cable bundle. A conducting asphalt is oftendesirable rather than a non-conducting insulating coat.This conductivity prevents build-ups of the current wave.

5.4.2.3 Braid Transparency. Transfer impedance was discussedpreviously in the analysis section, but for completeness, itwill be considered again in context of cable design. Ifthe wires of an unshielded bundle also interweave, then the"surface current" due to FIMP exposure gets transferredinside and all of the circuits share in its pickup. Abraided shield also behaves somewhat like this. Besideshaving lots of holes for field leakage, the braid wires divein and out. The wire-to-wire contact is not very effectiveand much of the surface current gets inside to radiate intothe internal cable zo*e. As an "average" example, an R';-8/Ucable has about a 10" current transfer ratio or transferimpedance, for submicrosecond pulses. Thus, at the end ofa j meter length exposed to a field pulse which induces10 amps peak current, there will be about a 50 volt sig-nal on the inside.

5.4.2.4 Outer Shield Construction. Here is additionalillustration of "good-bad" choices from the EIMP stand-point. Like normal enclosure shielding, the thicknessis not very critical. One finds that the emphasis is ratheron shielding "tightness". Again, we want no "cracks" if atall possible. One brand of "non-tightness" is that representedby a spiral-wound strip - even the double-layer variety.The trouble here is that each turn is actually a turn ina loosely coupled continuous mutual inductance. The contactresistance along the overlaps is too high to avoid somevoltage buildup per turn. This type of shield acts as apassably good coupler between external environment andinternal wires. Conduit, when properly installed bythreaded connection or welded joints, behaves very much likean additional solid-metal outer conductor. When neededfor other reasons, such as blast resistance or coderequirements, it is probably the cheapest in terms ofadded FEMP cost.

48~48

5.4.2.5 Quality Control. These problems can be compoundedby loose material specification, shoddy factory control, andlax acceptance criteria. Shielded cable deliveries shouldnot be judged for payment by stock clerks. Too often onefinds newly delivered cables in which the outer shield isbadly oxidized or even corroded. If the construction is suchthat good shielding depends on good internal contact, youare clearly not going to get it.

5.4.2.6 Outer Jacket or Sheath. The data on the valueof an insulated outer jacket on EMP/cable coupling remainsdebatable. There is some evidence that it can reduce theskin current in threat regions of direct field exposure andcoupling. On the other hand, contact of the outer shieldwith a conducting earth or the use of a conducting mediumhas also been shown to provide propagation damping for apulse originating elsewhere. It can also be argued that anouter jacket can "pinhole" through if the local field istoo high. The governing factor would appear to be durabilityto this end, lead sheathing or neoprene jackets seem toprovide lifetimes commensurate with normal useful systemslives. Of course, the fabrication and installation costfactors favor neoprene today. This is an area where furtherresolution of design considerations appears desirable.

5.4.3 Damping Schemes. EMP propagating in the directionof buried cable can cause a "traveling wave" voltage build-up. Here are some ideas on how this can be minimized.Besides "grounding the sheath", there are other steps whichcan be taken to alter and suppress the transmission offields along a cable. Suppose we look at a cable as thecentral conductor of a transmission whose outer conductor isvery far away. Any change in the transmission impedanceradically will produce reflections. Some of the energy willgo back where it came from; thus, the insertion of perpen-dicular conducting baffles. This has been successful inmany cases suppressing pulse interference in nuclear testcable systems. Anything we can do to enhance the fields ina "lossy" material (such as wet earth) will increase itsattenuative effects. This is embodied in the introductionof ferrite ring "field concentrators" around a cable. Ifthe ferrite is made lossy as well and is located at thebase of a baffle, an ultimate degree of feasible attenuationcan be achieved, as shown here. A "poor man's version" ofthe same thing consists of a spiral corrugated outer shieldwith interwoven high-P wire. At the moment, the aboveapproaches are in the "good ideas" category, and littleinformation exists on a quantitative basis.

49

5.4.4 Cable Terminal Treatment. After many years ofnoise and malfunction, systems designers learned to con-nect the outer shield of a cable to the metal can or frame-work of the equipment. The use of anodized shells formulti-pin connectors has created this problem. The onlyfeasible way to maintain a good "system ground" is to carrythe ground wire through one or several of the connectorpins. Of course, they did that poorly, too, by taking theconnection through a plug and socket pin, using a highinductance jumper. It isn't merely the insertion of theinductance which "kills" you, but the propagation discontinu-ity causes a pulse reflection at this point as wellalmost doubling the voltage induced across this interface.As suggested here, this injects a pulse into the signallines because of local capacitive effects.

5.4.5 Plugs and Sockets. The right way to do it is to usethe entire connector shell as the contact element. Thesocket should likewise make peripheral contact with theequipment shield or container. In effect, one wants toapproach the transmission effectiveness of UHF waveguidehardware, as far as the shield is concerned. Differencesin good overall shield design are often overridden by poorconnector design.

5.5 Conduit. When properly installed, conduit is easilythe best "outer shield" for cable systems. The principalproblems arise at segment contacts. Again, cleanliness andcareful assembly are essential. Rusted and corroded threadsand bushings will intiduce series impedances along theconduit's length. Welded joints are best, but expensive.Welding is almost the only dependable way to deal with theconduit terminals. Normal clamp rings make contact at onlya few points, at best. All too often, there are some looseones left behind. The conduit ends are particularly sensi-tive system points because here the exterior transmissionimpedance changes. Pulse currents flowing on the outersurface must be redistributed onto the equipment shield.If relative movements between exterior conduits and shieldedbuildings are expected (due to shocks or earth movement),the use of bellows, convoluted sections, or multipleknitted socks should be considered.

S.6 End Boxes. Cable terminal treatment by means oflumped-element "end boxes" is popular because it accentsthe illusion that "you're doing something about EMP".Usually, they contain elements intended to suppress EMPpickup by the cable wires themselves. This is fine, as

I s5

long as the final circuit reference level can be maintainedwith the internal zone. This then requires that the terminalbox be integral with the interzonal boundary, as shown here.Too often, such boxes are mounted on the wall with twoself-tapping screws, or are connected to the equipment canby means of a foot-long braided juniper. Cable end boxeshave also been effectively used for marginal retrofitcases, in which a more thorough "solution" would have beendramatically expensive.

5.7 Internal Treatment and Circuits. In some systems,such as communications, the cable cost is a large item,and the cable specification has to be detailed to achievemaximum service return. The internal conductors areusually specified according to the degree of environ-mental isolation which each service or function requires.The external shield then provides just the minimum neededby the least sensitive circuits. Frequently, advantageis taken of the specifications imposed by service perfor-mance requirements. Thus, a broadband coax requires asolid outer conductor and good lead connections for bestperformance.

5.7.1 Twisting and Shielding. For medium level sensorsand medium bandwidth circuits, adherence to zoning andcircuit reference criteria may suffice internally. Thismeans provision of separate return reference wires (indepen-dent two-wire circuits) which are fabricated as a twistedand shielded pair. Braided shielding is often sufficienthere. Broadly, these practices are generally commensuratewith requirements for intercircuit isolation ("crosstalk").

5.7.2 Balanced Isolation. As indicated before, if a systemlends itself to carrier techniques, then 30 to 40 dB may begained in protection value by using balanced cable circuitswith terminal isolation transformers. Conversely, therequirement on the built-in cable shielding may be thatmuch less severe. The advantage of the "carrier" approachresides in the case of obtaining balanced isolation transformerswhich operate over a "limited bandwidth". Extra filteringmay also be required to assure that the bandwidth is restrictedto the "limited bandwidth" of the transformer.

5.7.3 Terminations. It is good practice to considermulti-layer shields and common mode (phantom) circuitsas independent energy transmission lines and to terminatethem accordingly. As indicated before, precision in doingthis is not important - the main thing is to provide anenergy dissipater so that the pulse won't "rattle" back andforth.

51

5.7.4 Power Supply and Control Circuits. Some "connec-tions" nust be DC or non-carrier AC. If these can betreated by means of terminal filters and by high-leveloperation, then (together with carrier signal techniques)the EMP requirement on the total cable package can be aminimal one.

5.7.5 Hierarchies and Zone Loss. For short cable runs,one has to be somewhat meticulous about connecting succes-sive shield and conductor "layers" to enclosures andcircuits of corresponding zone level - no "criss-crossing".Obviously, one may otherwise end up by depositing high-level environmental "noise" (or EMIP) into low-level systemelements. It is also "bad practice" to "carry grounds" onthe outer skin in short runs, an easy temptation. For longruns, it was indicated that zonal identity tends to "getlost" in the damping due to radiation and medium coupling.At this time, this is more a matter of experience thantheory. By "long", is meant one maile or more.

5.8 Grounding is a complex topic, but can be over-simplified by dividing grounds into the interior andexterior grounds as indicated. To oversimplify somewhatfor clarity, there are:

a. Grounds/earths/buried conductors/rails/pipesoutside (exterior).

b. Grounds/reference nodes/buses/equipotential gridsinside (interior) for partially or well-shielded equipments.

"Grounds" are needed for almost "everything", but are oftena major source of MP pickup.

5.8.1 Ground Semantics. It is worthwhile reviewing brieflysome of the terminology in "grounding" or "earthing". Thetersm "ground" is often taken to mean a purposeful electricalconnection to an exterior buried conductor. It is alsoused to identify circuit connections to the "common" or"bus" side of a circuit or to chassis, racks or largeshields forming an "inside ground". Basically then, theorthodox/"outside ground" identifies an attempt to connect,in a field-significant way, to the large, but poor, rationalconductor which covers the earth's surface. This topic of"outside grounds" is probably the most ambiguous in thebusiness. The idea of "grounding" as a "good thing to do"originated decades ago when it was found that:

52

a. Power systems were less dangerous with controlledgrounds.

b. RF systems behaved more stably with controlledgrounds.

c. Antenna radiation improved with controlled grounds.

d. Lightning killed fewer people with controlledgrounds.

But there is a broad range of types of grounds and anequally broad range of quality.

5.8.2 Why is ElMP Concerned with Outside Grounds? Thereare at least three reasons for considering how EMP andexterior grounds interact. First, there are long wave-length threat components -- especially from surface bursts-- with which ground circuits can meaningfully couple.Second, system grounds are essential for any number of otherreasons; e.g., noise reduction, reference point, etc;hence, their EMP coupling is a germane issue. Third, ithas been pragmatically established that grounds make anoticeable difference - good and bad - in nuclear testinstrumentation. Of course, if a grounding system can beput to advantage in EMP control -- so much the better. Butthis may be in conflict with other grounding requirements;i.e., lightning, power, etc.

5.8.3 Ground Concept. The basic idea of a ground is toprovide an equipotential distribution between the dominantstructural members of a system and the surrounding naturalenvironment. That "equipotential" concept raises somequestions. Clearly, it is "perfectly" valid only for theideal case of:

a. Static fields.

b. Infinite conductivity.

c. No current flow.

Obviously then, if a "ground" seems to improve a real,non-static condition, it is because it does change theexternal field distribution. A current-ow must neces-sarily exist in the ground conductors in order to do this.Hence, real grounds always represent a departure from the

53

ideal. The point is that the external field distributionis a more tractable one in a "good ground" situation.Some of the mythology of the i3MP folklore concerns theimprovem~ent realized by avoiding "outside" ground loops bythe substitution of single-point straight cable runs. Ineither case, EMP can induce substantial amounts of EMPcurrent flow onto the cables or conductors. The outsideground loop does tend to enhance the lower frequency com-ponents and prolong the duration of large circulatingcurrents. The straight run tends to dump extra currentinto the interconnection point. In either case, the pickupis sensitive to angle-of-arrival, ground parameters (o andp) and many other details.

5.8.4 Ground Quality. The best way to evaluate the exter-ior ground concept is to point out the nature of itslimitations. Some systems could not realize any advantagefrom grounding. For instance, one can think of a configur-ation in which the interior zones are so well isolatedfrom the exterior that no external changes can be sensedinside. And, of course, there are the inapplicable caseslike missiles in-flightl Most systems are not that opaque.By proper design, the field and curre .t enhancements cangenerally be suppressed arising from exterior grounds.There are two basic limits where grounds are of littlevalue:

a. A ground connection so long that it introducesappreciable impedance in the ground circuit.

b. Dominant wavelengths so short that the system isphysically the larger.

5.8.4.1 Exterior Ground Resistance and Transient Impedance.These concepts are perhaps the least understood in con-temporary electronics engineering. To place them in theirproper perspective, it is essential to understand the electro-magnetic circumstances in a current-carrying conductor.Most of the iM energy exists in the field external to it.The energy content in the "skin current" on the conductor isusually a few percent. Thus, the effectiveness of a "ground"for transients or for RF in the LF to HF domain depends,in part, on the coupling of that external energy intothe earth and more significantly, its absorption. Bythis, we mean conversion of EM energy to heat (in the earth)by I R heating, of jourse. This is a very tiny number intypical cases 10 C rise. In many cases, the best choice

54

U Ai

might be to maintain a single-point exterior ground conceptconsistent with the position of the interior single-pointground for a shielded system. The power, lightning, safety,or crypto ground should be made as compact as possible,thus avoiding long conductor runs capable of collecting orenhancing EMP.

5.8.5 Inside Grounds. To understand why "inside" groundsare different, let us review a few facts. With a goodshield, magnetic field and coupling effects tend to dominate.Option: Use a good shield and minimize magnetic fieldcouplings and common impedance "voltage sources" by use of asingle-point ground. The shield "goodness" must be suchthat the wavelengths of the principal penetrating waveformsare always large compared to the longest ground cablelength.

5.8.5.1 Interior Grounding Systems. There are a widevariety of geometrical arrangements for "connecting toground". Figure 9 shows a typical ground system. Here,we identify several of the accepted connection geometries:

a. "Crow's Foot" or single-point. Probably the wisestchoice in a "bleak" situation, since it minimizes couplingin the ground connections proper.

b. Fishbone. The lower level (higher sensitivity)circuits sho uldusually be at the "far end", where "groundcurrents" are lower.

c. Multipoint. Note the opportunities for groundloops and common impedance IZ common voltage rises.

d. Floatinu Grounds. Often employed where a singlepoint ground is impractical and where a multipoint sys-tem could cause trouble. Here, each subsystem case assumesits own potential without ill effects provided that goodisolation (common-mode rejection) is realized.

5.8.5.2 Internal Grounds. It is very easy to get carriedaway to the point of providing good coupling between highenergy circuits and high sensitivity components by groundingeverything. The reference node design technique tries toensure that each electrical system is treated as a complete,independent circuit entity. Bonding between each level canthen be introduced selectively and carefully.

'55

aj!

-=a -

iziIIzI.

-J-

iii

56

5.9 Bonds.

5.9.1 Interzonal Bonds. As suggested here, the referencebond between successive zones should be approached with muchthe same criteria in mind as for any circuit connections.In particular, minimize BA. This means that the connectionshould be close to the main conductor runs and that thereshould preferably be just one of them. In some cases, the"bond" may be provided by massive and distributed struc-tural elements. This represents an "opposite extreme"which can be advantageously used. But then, it becomesexceptionally important that these extended contacts remainelectrically "good" -- no loosening of the bolts, or oxida-tion between the plates, etc.

5.9.1.1 Bond Straps. Bond straps or "jumpers" usuallyappear in a system as a result of faulty initial design.They can only be justified a priority, for conditions whereunusual flexibility or frequency disconnection are unavoid-able. The best way to connect things together is to bolt orweld them to each other and not use a bond strap at all.But if you must, Figure 10 shows what a bond strap lookslike in simplified mechanical and electrical forms. Acircuit analysis is presented in the handbook. Typicalgood bond straps have inductances in the order of 0.02 Hz.It is difficult to work effectively with much less thanthis, because then you might as well bolt the partstogether. Much more, and you're likely to get sparksbetween them.

5.9.2 Good Bond Strap Practices. Obviously, anythingwhich lowers the reactance of a bond connection improvesits performance. We see here the "good" directions. Ofcourse, ultimately, this would translate into a continuoussheet connection, and we would really have a piece ofshielding. Hence, there is an economic break point at which"more bonds" should really translate to "component redesign".

5.10 Protective Devices and Techniques. Up to now, pro-tection methods which, in one way or another, deal withthe coupling of the threat pulse environment to a systemand its circuits have been discussed. One might very wellask: Why was the discussion of filters, limiters, and soforth placed at the end? Often one hears that the "solution"to an EMP problem was a spark gap, a line filter, or someother lumped element placed at one critical point. Eachexpression tends to overlook the many other environmental

57;4~4

.4s

cLii

00

-iFFE

S8.

protection features which the system may have had builtinto it, intentionally or not. These made it possible toisolate the ultimate weak points to a number, type andsusceptibility, which could be successfully rectified by alimited and feasible lumped element approach. In general,protective devices and techniques alone could not protect atotally "naked" system at finite cost. This document willnot provide specific protective devices by manufacturerpart number, etc., but will provide a short introductioninto use, purpose, name, etc. Various references in theBibliography, such as 5, 6, 12, 18, 19, 28. and29 shouldbe consulted for more specific information on protectivedevices.

5.10.1 What is a Protective Device? Basically, it is a"lumped element" both in hardware and in software. Inhardware, it is most usually represented by a "black box"insertion in a circuit. But it can also be a mechanicalcontrivance inserted to interrupt a non-circuit conductivepath. In software, it can be represented bya functional insertion, such as a contingency command, or asensor-controlled branching in an operational sequence.In summary, it is an isolated object or action which isinserted specifically to counter an undesirable EMPconsequence.

5.10.2 Basic Protection Element Concept. This summarydefinition at once lends to a very basic specificationconcept for a protection element. An efficient protectionelement has a performance characteristic which is inverse tothe susceptibility of the hardware or function which it isintended to protect. One consequence of this is that suchan element must then be able to stop or absorb the excessenergy (or the erroneous message information) which wouldotherwise have reached the next system component. Usually,it would not do much good to protect a solid-state deviceby means of a second device of similar susceptibility.

5.10.3 Categorization. Some EMP protection schemes sug-gest that their designers were unaware of more appropriatehardware schemes. It thus seems worthwhile to categorizeand identify the many devices and techniques. The firsttwo are self-evident - hybrid devices combine voltage orcurrent limiting with filtering. Operational or functionalschemes involve various ways of interrupting system oper-ation and/or temporarily suppressing its sensitivity toreduce the threat consequences. Unconventional methods

59

include schemes for actuation or communication not requiringconductive paths or significant apertures.

5.10.4 Circuit Isolation. Passive and active lumped-elelent "boxes" are examples of isolation devices. Theycouple one circuit to another only in the signal regimeof interest. Limiters do this by restricting the rangeof linear transmission filters, by similarly limitingthe frequency domain. There are a number of useful passiveelem~ents besides filters, as will be discussed next.

5.10.5 Inductive Devices. Two inductive devices are particularlyuseful in comnon-mode suppression, as may be required oncable connections between equipment in two different EMzones. Both are extensively used in nuclear test instrumentation.In a bifilar choke, the push-pull or desired circuit pathscarried by the multi-conductor cable are only weakly coupledto the core, whereas the common-mode carried by the wholebundle is strongly coupled. Hence, the common-mode isstrongly discriminated against. The required series induc-tance evidently depends on the desired attenuation and onthe predominant frequency content. The balanced modetransformer similarly discriminates against common-modeenergy, but it is limited to HF application, of course.

5.10.6 Passive Devices - Filters. The most common passive,lumped-element device is the terminal filter. It isbasically a black-box with input and output connections forinsertion into an otherwise continuous two-wire circuit.Its insertion loss is chosen for least attenuation in thefrequency domain of normal circuit operation and maximumloss in the domain of maximum "noise" content. Note thatthe intercerted energy has to go somewhere. Maybe it isreflected back into the input system, to increase the EMPlevel there. Better that it should get "dumped" as heat inan internal filter resistance. This implies a preference for"lossy" filters, of course.

5.10.7 Butterworth Filters. Here is a typical workingexample of a filter analysis for a very severe exposuresituation. One might experience something as bad as thisfor a completely "naked" megawatt level power line.

5.10.8 Device Construction. The construction and instal-lation of a protective device is often as critical as itsdesign. If we think of a filter as a controlled RF barrier,

60

[ [ [['

then it is clear that its input and output must be isolatedfrom one another. A good filter (or other device) isusually constructed in three separate electromagnetic sec-tions; an output compartment, device compartment, and aninput department. Most frequently, filters and limitersoperate "against ground"; that is, the "return" side of theprotective element is well bonded internally to the filtercase. Good filter design and adjustment takes into accountwhatever mutual coupling may exist between input and outputwithin the central component compartment. This conventioncomes from the customary circuit practice of using "case" asthe reference node in small and medium size system elements,both for single-ended and balanced systems.

5.10.9 Device Installation. Obviously, the same care inisolation is called for in installation; much of the device'svalue is lost if the output side can "see" the input side.In the "right way", the filter case must make a tight periph-eral contact so that there is no "hair-line" aperture and sothat the common reference impedance is nearly zero. This isalso important if one is to obtain the benefits of thedesigner's and manufacturer's ratings. Evidently in suchinstallations, the internal EM zone makes a detour into thedevice's output compartment.

5.10.10 The EMP Box. In some installations, particularcare has been taken to isolate such entrance protectivedevices. This is the origin of the "EMP Room", sometimesostentatiously displayed as the "solution to EMP". Onolder, "unprotected" systems, one finds similar entrancespaces, simply labeled "cable termination vault". Whenproperly outfitted, these installations have value indecoupling the exterior from the interior environment and inreducing the secondary effects of non-linear operation ofthe protective devices themselves.

5.10.11 Active or Non-linear Devices. Partly in conse-quence of lightning and power-surge problems, a broadgamut of non-linear protective devices have evolved.Many of these are applicable to circuit EMP problems,perhaps with slight modification. The most serious defectin commercial protection devices is the penchant for using"pigtail" type connections between terminals and the pro-tection element itself. These generally present at least ashigh an impedance to an "EMP" as does the circuit itself.Many of these devices would be useful for EMP protection ifthey were simply "cleaned up" by using low-inductance bond

61.*1

straps and adequate connection contact areas especially tothe case (or "co;umon reference") side. In most cases, thesignal circuit should be taken through the box; the pro-tective device should not simply e sunted at a singleterm,|inal point.

5.10.12 Hybrids. This is the most favored "lumped-ele-ment" solution, in that is combines the better featuresof active and passive elements. The filter element sup-presses "hash" below the breakdown level, as well assuppressing hash generated by the active device itself.The series impedance preceding the surge arrestor is anecessary component to assure appropriate limiting; in somecases, the surge impedance of the transmission line cansuffice.

5.10.13 Types of Non-linear D)evices. This representsa listing of "possible" devices for EMP application. Oneproblem with most non-linear mechanisms is "hang-up" orhysteresis in the activation/deactivation cycle. Thus, itis difficult to use them in circuits in which normal oper-ation may involve levels approaching "breakdown". In somecases, it is impossible to avoid some degree of "cold restart"capacity such as might be provided by electromechanicalrelays or 1y "crowbar dumping". This applies particularlyto high power level systems, such as utility distributionor radio transmitter outputs.

5.10.13.1 Consequences of Non-linear Operation. Non-linear devices are not unmixed blessings. We alreadyindicated under filters that the EMP energy has to go somewhere.This remains true for active elements as well. Furthermore,one is also faced with the feature that the switchingoperation itself can be a source of unwanted EM energy(e.g., RFI). This is particularly true if the associatedcircuits contain significant B1 energy in normal operation.When the device switches, it must inevitably cause somechange in effective circuit impedance and, hence, in operativecurrent distribution. In addition, the switching functionmay generate a spurious pulse in the circuit itself. Thisis particularly possible if the switching occurs on a timescale short compared to that of the normal operationalsignals in the system; e.g., on the fast-rise "front" of aninduced lEMP signal. This is one of the strongest reasonsfor using "hybrid" lumped elements.

62

5.10.13.2 Spark Gaps and Gas Diodes. These both dependon initiating conductive breakdown in a gap. Spark gapsare bipolar in operation, have low voltage drop when conducting,and are simple and easy to make. But they do not extinguishwithout removal of almost all power, and their breakdowncharacteristic is such as to generate considerable 11F"hash" in their vicinity and in the connected circuits.Gas diodes operate as smaller voltages and "turn on" lessnoisily, but they cannot carry high currents for long andtend to be capricious as to long-term reliability.

5.10.13.3 Zener and Silicon Diodes. These are generallysmaller, lower power devices. They operate effectivelyin the voltage-current range of solid state circuitry,so that they are extensively used for such circuit pro-tection. They are voltage-limiting in action (rather thanvoltage-reducing). Silicon diodes "clip" more effectivelytheir "plateau" is flatter. Their operating voltage isgenerally low - a few volts - and the introduction of "hold-off"bias can be an inconvenience. They are generally high-capacity devices, so that there are limits as to the circuitfrequency range of applicability. Also, the semi-conductordevices have definite limits on the joule energy handlingcapability.

5.10.13.4 Thyrites. This may be used for "brute force"problems. It can be used for good transient response.It is basically a non-linear resistance with unusuallyhigh power-dissipation capacity. It can be made "simple"by virtue of using circuit impedance as part of the limitingmechanism, but it is not an "absolute" limiter - it only"rounds off" a transie- peak.

5.10.14 Fast Relays. These relays operate in less thanone millisecond. Their principal value lies in inter-rupting protection circuits in order to limit the energydissipated in the faster protection devices, and in orderto initiate restoration to normality from a breakdowncondition.

5.10.15 Crowbar Circuits. In these systems, a high-power rating device is operated by a subsidiary sensing/trigger circuit. Thyratrons, ignitrons, and spark gapshave been used for the "crowbar". Sensing can come fromthe circuit itself, from a "threat" sensor (see last topic),or from an auxiliary breakdown device (e.g., corona opticalsensor). Crowbar circuits are often used to activate thenormal system protection interlocks. For example, EMP

63

could "fire" a spark gap or cause an arc-over in the trans-mitter output. This arc, if not extinguished, could causeexessive plate dissipation in the output tubes. In thiscase, a thyratron can be "fixed" across the transmitter DCsupply to activate the circuit breakers. The thyratron isactivated by a corona-sensing photocell near the spark gapor, better yet, by an impedance sensing circuit.

5.11 Circumvention Techniques. Some EMP problems arephysically sufficiently formidable and "hard-nosed" hardwaretreatment is simply impractical. This can be particularytrue in retrofit - perhaps for no other reason than thecomplications involved in removing system elements fromoperational status. The basic principle involved in "cir-cumvention" is to reduce the time period of high vulner-ability so as to reduce the probability of coincidence withthreat exposure. Obviously, human intervention is (almost)out of the question. (The only conditional situation hereapplies to systems intended for post-strike activation.)There are two broad classes:

a. Non-threat-specific or duty-cycle techniques.

b. Threat-specific or nullification techniques.

5.11.1 System Constraints. If we "gate down" a systemin real time, there is at once an implication that theoperational sequence execution would take place on acomparable timescale and with appropriate bandwidth.Only the more modern and sophisticated systems have suchcapabilities (i.e., 10 usec stepping time). But, suchsystems are also open to a number of protection responseoptions. For instance, th . system response can be pro-grammed to depend on where in the sequence the threatappears. It can overlook the threat if it is in a relativelyinvulnerable mode. It can stop and restart from some pre-viously determined early stage or cancel a number of pre-vious commands. It can similarly pause or hold, test forstatus validity, and start up again. Older or more primi-tive systems generally cannot be desensitized "in time".Usually, one must assume error or interruption, when athreshold field is reached, and simply restart the sequence.(This presumes that the system is hard enough to avoidpermanent damage, of course). In some cases, one may havethe option of programming a separate sequence validity testwhich can negate or enable the mission sequence at somelater stage.

64

.... I

5.11.2 Non-Threat-Specific Schemes. Duty-cycle schemesare generally permissible when the exact threat responsetime is not critical. If a particular system step requires11 Psec to execute, but may be done anytime within 10 Psec,then one may gain a reduction factor of 100 in threatcoincidence probability by suitable cycle suppression.Both random and synchronous schemes have been considered.The synchronous scheme lends itself to certain forms ofbandwidth reduction as well. A variety of gating andswitching techniques can be applied for disabling circuitinputs during the "off" periods. Redundant message trans-fer is another alternative.

5.11.3 Threat Specific Schemes. An active nuclear threatmalay be sensed in a number of ways. Let us confine atten-tion to "prompt-spike detection". The basic reason thatthis works for EMP is that the waveform peak inside asystem is generally much broader and, hence, later thanoutside. In principle, one can use the exterior-sensedsignal to vate down" the execution sequence before theinternal environment reaches error or interruption levels.The biggest problem with this scheme lies in "false triggers".Experience indicates that it is almost essential to coupletwo different prompt sensors in coincidence in order toavoid almost continuous system inhibition due to non-nuclearnoise. By "different", we really mean different, such asan EMP antenna and a photoelectric unit.

5.11.4 Non-electrical Schemes. Some relatively simplemechanical schemes have been applied to zone decouplingproblems such as non-conducting shafts in M-G sets andinsulated solenoid actuating switch.

5.12 Economics. Finally, a word about EMP hardware costs.The lack of meaningful improvements in EMP hardwarePiethods is partly a consequence of its frightening economics.This fear of costs in itself engenders higher costs, sinceit develops so few reliable "cost effectiveness" experiences.It is fairly clear that meticulous adherence to appropriateknown hardware techniques adds noticeably to system costs.At present, there seems to be no reliable avenue for safecompromise. Most economical and effective protection isrealized if the hardening effort is an integral part ofthe original system design.

65

BIBLIOGRAPHY

1. DNA EMP Awareness Course Notes, DNA 2772T, Second

Edition, ITT Research Institute, Chicago IL, Sep 73

2. "EMP Protection for Emergency Operating Centers", TR61ADefense Civil Preparedness Agency, Washington DC, Jul 72

3. "EMP Protection for AM Radio Broadcast Stations", TR61CDefense Civil Preparedness Agency, Washington DC, May 72

4. "Nuclear Electromagnetic Pulse Protection", Army TM855-5,Washington DC, Feb 74

S. Vance, E. F., Design Guidelines for the Treatment ofPenetrations Enterin Communicato-FsFcilities;Stanford ReseiarchTInstitute, 1975, [A U 7U 76

6. "DNA EMP Handbook" (U), DNA2114, Volumes 1 through 4,Defense Nuclear Agency, Washington DC (S)

7. Vance, E. F., "Electromagnetic Pulse Handbook forElectric Power Systems", SRT Project 2709, Stan-ford Research Institute

8. Clark, Don, "EMP Protective Systems", Office of CivilDefense, PEM Note 9, published by Lawrence LivermoreLaboratories, Livermore CA for DNA

9. "NEMP Protection Inspection Guide for Safeguard TSESystems and Equipment", HNDDSP-71-42-SE, US Corp ofEngineers, Huntsville AL 35807, May 71

10. EMP Threat and Protective Measures, TR61, DefenseCivil Preparedness Agency, Washington DC, 1970

11. Lasiter, H. A., et al; Nuclear EMP Pulse EffectsDesign Parameters for Protective Shelters, Naval CivilEngineering Lab, Port Hueneme CA, June 1968

12. Lasiter,.H. A., et al; Nuclear EMP Pulse ProtectiveMeasures Applied to a Typical Communications ShelterNaval Civil Engineering Laboratory, Port Hueneme CA,April 1970

13. Peele, W. D, et al; "EMP Protection Study", FAA-RD72-68,RADC, Rome NY, June 1972

66

14. Bridges, J. E., et al; EMP Directory for ShelterDesign, PEM Note 2, Lawrence Livermore Laboratory,Livermore CA, December 1971

15. Wouters, L. F., "EMP Protection Engineering - TheNeed for Documentation" PEM-1, L.L.L., Livermore CA,March 1972

16. Warner, G. L. and Doskocil, A. C., "EMP Hardeningfor Sam D", PEM-25, L.L.L., Livermore CA, Sept 1973

17. "EMP Preferred Test Procedures", ITT ResearchInstitute for DNA, Washington DC, Aug 1974, AD787482

18. "EMP Electronic Design Handbook", AFWL TR 74-58,AFWL/ELA, Kirtland AFB NM 87117, April 1973

19. "EMP Generated Waveform Specification of Subsystemand Circuits for US Army Missile Command" byRaytheon Company, Bedford MA, 1972

20. "Protection Instructions, Hardness Program - EMP"Volume I, HNDDSP-72-145-ED-R, US Army Corps ofEngineers, Huntsville AL, December 1974

21. Nelson, D. B., "A Program to Counter the Effectsof Nuclear EMP in Commercial Power Systems ORNLTM 3552, Oak Ridge National Lab, October 1672

22. Carter, J. M., et al; "Aeronautical Systems EMPTechnology Review", AFWL TR 73-11, Boeing Company,July 1973

23. Davidson, G. G., et al; "Characterization ofEMP Protection Devices, Army Electronics Command,Ft Monmouth NJ, July 1973

24. Keyser, R. C., "Satellite Hardening Techniques,Evaluation and Test Definition", IRT Corporation,AFWL TR 74-3281, September 1975

25. Vance, Edward F., "EMP Handbook for Electric PowerSystems", Stanford Research Institute prepared forDNA, 4 February 1975, ADA009228

26. "EMP Hardening of GFE", AFWL TR 74-61, AFWL KirtlandAFB NM 87117; July 1973, AD918276

67

.. ... . . _ . . . - - ... .. . a = _ i . . - - - --

27. "IEMP lHandbook for Missiles and Aircraft in Flight",Sandia Labs for AFWL, AI'WL 'rR 73-68.

28. "Analysis of Failure of Electronic Circuits fromEMP Induced Signals" Harry Diamond Lab HDL-TR-1615,August 1973

29. "EMP Electronic Analysis Handbook" prepared by BoeingCompany for AFWL; WL TR 74-59, 1974

30. Lunn T. C., "Techniques to Determine SubsystemLevel EMP Susceptibility", PEM 36, Lawrence Liver-more Laboratory, Livermore CA, 1974

31. Whitson, A. L., "DCA HEMP Hardness CertificationMethodology Status", Stanford Research Institute,February 1974

32. Story, Kenneth E and Row, Ronald V., "Conducted EMPfrom a Commercial Prime Power Network", PEM 48,Lawrence Livermore Laboratory, Livermore CA

33. "Development of Design Criteria Relating to NEMPEffects on Power Systems", General Electric Company,Pittsfield MA, for Office of Chief of Engineers,Depot of the Army, December 1971

34. "Development of HEMP Assessment Methodology forSatellite Terminals", Intelcom Rad Tech for DefenseNuclear Agency, DNA 3553F-2

35. Durgin, D. L. and Brown, R.M. "Key SuppressionDevice, Parameters for EMP Hardening", Braddock,Dunn and McDonald for Lawrence Livermore Laboratory,PEM 38, March 1975

36. Baird, James K., et al; "Effects of EMP on the Super-visory Control Equipment of a Power System", OakRidge National Laboratory, Oak Ridge TN, October 1973

37. Morgan, G. E., et al; "Internal EMP and SystemGenerated EMP" Antonetics Group, Rockwell Inter-national, June 1974

38. Gould, Ken E., "The High Altitude NEMP - An Intro-duction for Managers of Federal TelecommunicationSystems" General Electric Company for DefenseNuclear Agency, DAS/AC TN 74-2, June 1974

68

& 39. Clark, 0. Melville Winters, Richard P., "Feasi-bility Study for EMP Terminal Protection" preparedby General Semiconductor Industries, Inc., 1973

40. "Grounding, Bonding, and Shielding Practices andProcedures for Electronic Equipment and Facilities",Proposed MIL-HDBK 419.

[-

1'

69

DISTRIBUTION LIST

Addre ssee Number of Copies

Defense Documentation CenterCameron StationA lexandria VA 22314 .. .. .. ... . . .. . . S

NCA/EPE..... .. .. .. .. .. . . ..

E CA/EPE . . . . . . . . . . . . . . . . . . . . . 2

P CA/XME . . . .. .. .. ... . . *. .. .. .... 2

AFCS/OA . . .* * . . . .* 2

A FCS/EPT . . . . . . . . . .* . . . S

1842 EEG/DCIS . . . . . . . . . . . . . . . . . 2

1842 EEG/DCTM . . . . . . . . . . . . . . . . 2