INDUSTRIAL APPARATUS CONSULTANTS LIMITED (62 Marylebone High Steet, London, W1U SHY TEL: 020 7486 6474 FAX: 020 7487 2757 A Guide to the Use of Spark Gaps for Electromagnetic Pulse (EMP) ProtectionAbout the AuthorsTable of Contents Foreword Author's Note Abstract About the Authors Introduction. , eee 1.0 Spatk Gap Protective Devices... soos vital 1.1 History and Development : 1 1,2 Examples of Spark Gap Applications ... ane . ; 3 1,3 Spark Gap Operation 14 Types of Available Gas Tube Spark Gap Protector 1.5 Performance Information at 2.0. EMP Considerations . 2.1 Background 2.2 EMP Environments 2.3 Coupling Modes .......-- ‘24 Hardening Techniques ...- 3.0 Applications Information 3.1 Introduction... 3 3.2 Advantages and Disadvantages of: Soak Gaps. 3.3 DC Circuit Application .. . 34 AC Circuit Application q 3.5 Application of Gaps to Communication Circuits. i 516 Application of Special Devices, the Thee Electrode Spark ee ees 3.7 Installation Considerations .....2...22.0+ ‘i 3.8 Testing and Hardness Assurance . 3.9 Illustrations of EMP Applications [Appendix A ~ Spark Gap Current» Voltage Characteristic Curves. ‘Appendix B — Waveshape Designation for an Impulse Appendix C~ Description of an EMP Spark Gap Speifcation « Appendix D — About Joslyn ‘ 7 References ...-List of Illustrations Figure | — Volt-Time Curves, 6 Figure 2 ~ Spark Gap Response 6 Figure 3 — Voltage Waveforms 8 Figure 4 — Surge Current Waveform . : 8 Figure § ~ Surge Voltage for Gap with Current Limiting Resistor .... . 9 Figure 6 — Surge Current Waveform for Gap with and without Series Resistor promo aae Figure 7 — DC Holdover Curve . eric Figure 8 ~ Follow Current a 2 Figure 9— Photograph of Basic Two-Electrodes Low-Energy Gaps. -... ss sss, 13 Figure 10 — Cross Section of Basic MSP Gap . : as Figure 11 ~ Photograph of Two Electrode High-Energy Protectors Hens Figure 13 — Comparison of Two Separate Two Electrode Gaps with One Thiee-Electrode, Common-Chamber Gap... seen Figure 14 — Development of Common Mode Voltage, Two Single Spark Gaps Versus One Three-Electrode, Spark Gap : mace Figure 15 — Photograph of Low-Eneigy Thiee-Electrode Gaps 219 Figure 16 ~ Cross Seetion of a Low-Energy Common Chamber Three Electrode 20 Figure 17 — Photograph of High-Energy Three-Electrode Gaps igure 18 — Cross Section of a High-Energy Three-Electrode Gap igure 19 ~ Photograph of Hybrid Protectors... Figure 20 ~ Cross Section of Typical Hybrid Protector Figure 21 — Two Electrode Gap with Series Resistor Figure 22 — Three Electrode Gap with Series Resistors... Figure 23 ~ Slow Rise Impulse Test Ciruit Figure 24 — Fast Rise Impulse Test Circuit “ Figure 25 — Volt-Time Curves for Two-Electrode Low-Energy Protectors 28 Figure 26 — Volt-Time Curves for Two-Electrode High-Energy Protectors ..........00.+ 30 Figure 27 ~ Volt-Time Curves for Three-Electrode Low-Energy Protectors al Figure 28 ~ Volt-Time Curves for Three-Electrode High-Energy Protectors 234 Figure 30 ~ Simple Comparison of EMP Regions aaEHHeeSMatS + al Figure 31 ~ The Compton Effect : 39 Figure 32 ~ Example of DC Acutal and Equivalent Circuit sa Figure 33 ~ Example of Spark Gap Installed Next to Load 55 Figure 34 ~ Surge Current Flow When Gap is Considered a Short Circuit 3s Figure 36 ~ Equivalent Circuit for Extinguishing Criteria . 37 Figure 37 ~ Location of Added Resistance, Radded . em cee SB Figure 38 — Spark Gap-Resistor Combinations Added in Parallel 59 Figure 39 — The Effect of Adding Spark Gaps in Series 60 Figure 40 ~ DC Power Circuits - Considerations for Spark Gap Selection 61seems COA NNTP RY Figure 41 — AC Power Circuits - Considerations for Spark Gap Selection 3 63 Figure 42 — Communications Circuits - Considerations for Spark Gap Selection 4 Figure 43 ~ Typical Application of Special Devices 65 Figure 44 ~ Spark Gap Installation Diagrain : 68 Figure 45 ~ Photograph of Aircraft Protectors 68 Figuie 46 — Representative Circuit Applications 9 Figure 47 ~ Exponential Decay Current Pulse 10 Figure 48 ae uss eal Figuce 49 2 Figure 50... i n Figure $1 ~ Test Cireuit for 5002-01 n Figure $2 ~ Test Circuit for FS001-01 z B Figure v1 ~ VoltAmp Characteristic fora Typical Spark Gap 78 Figure B-1 ~ Wave Shape Designation of an Impulse 7 Figure C-1 ~ Response Time Definition 5 stipes 7” Figure C-2 ~ Pulse Potential Waveform pias ges List of Tables ‘Table 1 — Non-EMP Transient Sources... ere 3 ‘Table 2— MSP Spark Gap Specifications : ees ees was pelt Table 3 — Specifications for High Energy Protector With Integeal Current Limiting Resistor ...2....022000005 16 ‘Table 4 — Typical Specification for a Low Energy Common Chamber Three Electrode Spark Gap... emo all Table 5 ~ Specifications - 40 Coulomb Three Element Spark Gap aac a ‘Table 6 — Typical Specification for Low Surge Protectors Having High Series Resistance ...- ae os a 28) Table 7 — Typical Specification for Multiple Packaged Low Surge Protectors Having High Series Resistance... danas ‘Table 8 — Comparison Chart of Common Protection Dé Designed for Use in EMP Applications... a st ‘able 9— Operating Cteuis for Aiteratt EMP Spark Gap Appliation 5 68 ‘Table 10 ~ Operating Characteristics for Aircraft EMP Spark Gaps weeeesTO Table 11 — Joslyn Spark Gaps Used in a Missile System ’ n ‘Table 12 — Joslyn High Current, Three-Element EMP Protector... -- 73 viiIntroduction Now that the existence and importance of electromagnetic pulse (EMP) effects have been recognized, the task of providing effective EMP protection for existing and future systems must be dealt with. However, EMP technology is a difficult and specialized field. Engineers and technicians who are not EMP experts need information and training in order to work on the design and implementation of EMP-protected systems. The use of spark-gap protective devices is one approach that may be used effectively in some situations. However, spark-gap technology is also a specialized field and the number of competent engineers and technicians is limited. This book sims to introduce those confronted with EMP.protection problems to both subjects. ‘The information in the book is presented in three sections. Section 1 Is about spark gaps, how they ‘were developed, how they are used, how they work, the different types available and their test performance, Section 2 is about EMP, the EMP environments, coupling modes and hardening techniques. Section 3 tells when and how to use spark gap protectors for EMP applications and gives examples of current applications. Appendices are included which provide added information on specific points of spark gap application and a short description of Joslyn1.0 Spark Gap Protective Devices 1.1 History and Development In 1752, Benjamin Franklin encountered the principle of the spark gap, when during his electric kite experiment he placed his knuckle near @ key tied to the kite string. Suddenly 2 spark jumped from the key to his finger. At that instant he showed that thundetheads are electrically charged and simultaneously established his reputation as one of the most foolhardy experimenters in the history of science. ‘Sparks associated with static electricity have long been a phenomenon of interest. Dr, Franklin's discovery, however, was the forerunner for significant experiments and developments in electricity and electrical protection. For example, soon after his electric kite experiments, Dr. Franklin devised the lightning rod. Also, after several fatal accidents during electrical experiments, methods began 10 bbe developed for the protection of experimentalists and equipment, “The earliest reference to a spark gap protector that we could find is in an 1859 Yale University textbook on Natural Philosophy (Reference 1). This early handbook in applied physics describes Dr. Franklin's experiment and his invention of the lightning rod (with complete instructions for lightning tod construction), and also gves the specifications fora spark gap for use in electric kite experiments. Such experiments were described as using a metallic kite string connected to a post by alength of silk cord, with atrangements for charging Leyden jars. To provide protection the metalic string was connected to an insulated conductor, and a brass ball, firmly connected to the earth or water, was located within two inches of the conductor. When the electricity on the kite, the metalic string and the conductor became sufficiently intense 10 arc the two inch gap it was conducted safely to ground (eal ground in this cas). ‘The British chemist Humphrey Davy is credited as being the fist to seriously study the nature of the electric are produced between two' conductors. In 1808, Davy discovered that the gases in air serve as conductors — the are develops because the electric potential difference between the electrodes reaches an intensity sufficient to break down or “ionize” the air (or other gas) between, and the ionized gas then provides a conducting channel for an electric current, producing the “flame” or at. ‘As experimentation and use of electricity expanded, the spark gap came into common use for protection of circuits and equipment. Early spark gap technology involved the use of closely spaced ‘brass balls, but the carbon block spark gap eventually became more common because of its simplicity and versatility. Characteristics of modem carbon block protection devices are covered in detail in Section 1.3 of Reference 2.Spark gaps operating in ambient sir, essentially an uncontrolled environment, are erratic. This led to the development of techniques for enclosing the gap, and virtually all spark gap devices in use today are hermetically sealed. Over the course of their development, spark gap electrodes have been made from almost every conducting material available, and the insulation material between the electrodes (the gap) has ranged from a vaccum to a variety of gaseous, solid, and liquid materials at different pressures or densities. Also, various means have been used 10 extinguish and control the are such as current limitation and magnetic blow-out. During the 1950's the design of heavy-duty spark gaps was refined to provide protection against repeated direct lightning strokes. Through the use of techniques that move the point of contact of the arc around the electrode surface, higher current capacity and longer electrode life have been achieved for given sizes of spark gaps, An example of this type of spark gap is discussed in Section 1.4 under two-electrode, high-energy protectors. During the past decade, spark-gap technology has grown rapidly, largely because of the semiconductor. Semiconductor circuits are much more susceptible to transient damage than their ‘vacuum tube predecessors, and the need for spark gaps having greater precision and faster response thas resulted in the development of such spark gaps. Refinements in the selection of spark-gap-electrode configurations and coating materials, combined with experimentation in varieties of fill gases and radioactive iostope prompting, have produced devices with lower breakdown voltage ratings, faster response times, greater repeatability of breakdown and longer life characteristics than spark gaps used less than ten years ago, Sealing techniques have improved, particularly ceramic to metal bonds, which provide better long term control of gas pressure, a variable critical to spark gap breakdown, Conversion within the telephone industry to sealed, gastilled protector tubes from earlier carbon block spark gaps has provided the stimulus for research into improvements in both spark gap performance and volume produetion techniques. ‘The specialized requirements of the EMP conimunity have further advanced spark gap teclnology, particularly in the area of response times, Specifications requiring voltage breakdowns in the nanosecond region have become common and, where significant energy is involved, spark gaps are being used to provide the necessary protection. Environments associated with nuclear weapons effects have required accessory packaging and testing. onal capability during and following nuclear attack. Due to this need, to provide assurance of opera spatk gaps devices have been designed to meet specified shock and vibrational levels far beyond the requirements of commercial systems1.2 Examples of Spark Gap Applications In examining Joslyn applications engineering data we discovered that many of the same basic spark nps used for protection from EMP are also used for protection from other transient sources and vice verse. The main difference is in the protective packaging, mounting schemes, and test and inspection requirements, One of the things of interest in this respect were the types of transients other than EMP for which spark gaps are used, Since EMP protectors must also be capable of protect- ing equipment from non-EMP transients, we have listed some of these other transients in Table 1 Table 1 Non-EMP Transient Sources Natural Man-Made Lightning Switching (1) Direct or indirect strike | Arc welding (2) Conducted Contact with foreign circuits Aurora Loss of neutral-to-ground connection Corona Capacitive discharge of long cables Electrostatic Conduction, induction, radiation Radistion Dynamite explosion Ionization Fluorescent lights Motor starting Fuse blowing Breakers opening Corona Electrostatic Reflection waves Generator overspeed Generator hunting Harmonie overvoltage ‘System faults Restriking of breakers, arresters, | cutouts We were also interested in some of the non-EMP circuit applications for which spark gaps have been used. These are summarized next, EMP applications are discussed in Section 3. A Summary on Non-EMP Circuit Applications DC power protectors are used to protect the DC power input to the following devices: DC power used in relaying and control in AC power companies; in various circuits in the electronic field; inrailroad equipment; microwave stations using DC under AC/MG; and DC powered missile systems. AC protectors are used on the AC power input to installations of: missile sites; microwave st (both straight AC and the charger-battery type); solid-state control devices used by power companies; substations and remotely controlled devices; computers; and instrumentation in processing plants. Specific applications are: solid-state equipment, rectifiers, relay coils, 60-cycle powered instrumenta- tion, power supplies, standby power switching circuits, 60cycle input circuits for microwave stations, power companies, telephone systems, pipelines, railroads, and communications. In the signal, communication, and data link line spark gaps and arresters containing spark gaps are supplied to protect co-axial, balanced and unbalanced, and open wire lines. They are used to protect the input and some output circuits of the following equipment: cartier repeaters (power company, RRY, industrial); hotbox detectors (RR); CTC (Centralized Traffic Control) bungalows (R-R.); track overlay equipment (R.R.); wire overlay equipment (R.R.); and any data line or other “pair” input to microwave stations (R.R., power company, industrial, computer, facsimile). 1.3 Spark Gap Operation Basic Operation (Carbon Blocks vs. Gas Tubes) In general, the term spark gap refers to any device that forms an electrical arc between two electrodes ‘when the voltage across the electrodes is sufficiently large. The formation of an electric arc isa very complicated phenomenon that is not yet completely understood. (The lightning stroke is a ‘well-known example of an electric arc.) Are formation depends on the presence of a very high level clectric field, which tends to stip electrons from the positively charged nuclei of the atoms of the gas between the electrodes. Under the right conditions these electrons collide with other atomic electrons and release them from their atomie bonds, in a cascading fashion. The net result is a conductive path along which large numbers of electrons flow. This flow of electrons is an electric current. The whole process is much more complicated than indicated by this brief description; many parameters must be considered in determining the details of the electric breakdown. Scientific attempts to describe the breakdown process are far from complete, and the construction of reliable gaps still requires equal amounts of scientific knowledge and practical experience. AA primary group of parameters to be considered are (1) the voltage across the electrodes, (2) the time ‘ate of change of voltage, (3) the shape of the electrode, and (4) the space (or distance) between the electeodes, All these factors determine the electric field between various pairs of points on the electcodes, and the level of this electric field determines how fast electrons are extracted from the electrode surfaces. AA second group of important parameters is related to the properties of the medium between the gas composition, pressure, impurities, trace elements, ionizing radiation, ete. All of electrodes *Railroadneers emmnemmmcemses EERE UTI OC these elements determine just how, when, and where breakdown will occur. Several different types of protective devices can be included in the category of spark gaps. An older devie protector. This device uses closely spaced, flat carbon electrodes that discharge in air. Because these devices operate in air, their properties depend somewhat on local sit properties. The DC breekdown voltage for electrode spacing is fairly high (about 600 volts), and close electrode spacing may tend to permanently short out the gap oF cause noisy operation following one or more discharges. found in many communication systems for lightning protection purposes, is the carbon block AA second category of spark gap protectors is the gas tube, Here, the electrodes are enclosed in some type of housing, and both the gas mixture and gas pressure between the electrodes can be varied to control the breakdown characteristics of the tube. This capability gives the designer of gas tubes a ‘number of options, Its experimentally known that the breakdown potential of a gap depends on the product of the gas pressure and the electrode spacing (Paschen's Law) for a given gas mixture, electrode shape and material. Consequently, by varying either the electrode spacing or the gas ‘pressure, specific breakdown characteristics can be achieved. Gas mixtures can also be varied and trace amounts of radioactive isotopes can be added to ~pre-ionize” the gas to help stabilize the DC breakdown voltage or improve the response of the tube. Gas tube devices are much more versatile than carbon block protectors and can be designed to meet most of the fast response time requirements of EMP protection. Any further reference to spark gaps will be directed mainly at the gas tube type. Description of Operating Characteristics ‘As mentioned in the previous section, many parameters determine the breakdown characteristics of a spark gap. For a given device, a number of these parameters are predetermined by the manufacturer (eg. gas mixture and pressure, electrode design, etc.). However, the design engineer needs to understand what factors determine the spark gap's response once itis placed in a cirouit and how these factors are characterized. One factor which greatly affects the response of a spark gap is the nature of the voltage waveform applied across its terminals. This property can be described in terms of a volt-time curve, shown in Figure 1. On such a curve any straight line through the origin corresponds to a constant rate of change of voltage with respect to time. The slope of the line denotes the magnitude of this rate of cchange.* The point at which a given spark gap breaks down depends on the time rate of change of the applied voltage waveform. The breakdown voltage point of a given spark gap can be plotted as a point fon each linear ramp waveform (the straight lines on the volt-time curve). When all such points are plotted a curve results which is the characteristic breakdown voltage curve for the specific device type. * Sometimes volt-time curves are plotted on logdog or loglinear graphs. Constant slope curves on logdinear graphs then do not appear as straight lines, but the basic idea isthe same,w VOLTAGE VOLTAGE (V) CURVES OF V/T = CONSTANT DC BREAKDOWN VOLTAGE Figure 1 — Volt-Time Curves It is very important to understand that a spark gap witn a very fast response time provides better protection not only by clamping at lower voltages, but also by allowing less er to the load to be protected. nergy to be conducted Figure 2 illustrates this point. Applying a voltage wave with the same rate of rise to a fast response spark gap results in fast breakdown; ie. gap A fires at taV, Gap B, being slower, will ake longer (tp) to respond, and will allow the voltage across it to increase to Vp. The crosshatched area of tp = is proportional to the additional exposure a sensitive load may have to encounter. SURGE VOLTAGE WAVE GAP B FIRES WITHOUT SPARK GAP apuirines WITH SPARK GAP Va Figure 2 ~ Spark Gap Response time ts) CT eaeDC Breakdown Voltage Note that the volt-time curves of Figure 1 tend to flatten out and approach a horizontal fine at the right hand side of the graph. This area of the graph corresponds to very low values of dv/at being applied to the gap. The voltage being approached is known as the DC breakdown voltage of the gap. ‘A given general family of spark gaps can usually be divided into various members which are characterized by their DC breakdown voltages. The DC breakdown voltage is not of much use in itself, since usually for EMP protection a very fast rising voltage pulse is of prime concern. The engineer needs to evaluate the volt/time curve for very steep volt/time slopes. The DC breakdown voltage is useful, however, for characterizing gaps and checking their condition. The DC breakdown voltage is easily measured and can be checked to assure that a spark gap has not changed its properties ‘on the shelf or with use. In applying spark gaps, this DC breakdown is also useful, for a gap cannot be employed in a circuit in which the peak AC or DC voltage exceeds the DC breakdown voltage of the ap. ‘The area on the volt/time graph of most interest for protection against very fast EMP transients is near the left edge or very steep constant v/t curves. This i also the area where spark gap operation is least understood. Spark gaps can be made to react very quickly but they are limited in response time bby the transit time needed for electrons to cross the gap. In addition, any inductance of the electrodes, leads, or series resistors will increase the response time still further. Consider how a spark gap affects a voltage transient applied across it, Figure 3 shows the voltage waveform that would be present without a spark gap and the waveform resulting from the gap's breakdown, At time Tg the gap begins to arc. The exact time depends on the volt/time slope of the applied voltage waveform and the specific volt/time curve of the gap being used. After the spark gap has fired, the voltage across the gap falls to a low level called the arc voltage and stays there for the duration of the applied pulse, The arc voltage is one of the built-in characteristics of a spark gap. Now consider how one finds the current that flows through the gap. To do this one must characterize the EMP induced signal by an equivalent voltage source with some series source impedance. No ccurtent will flow until the gap begins to arc at time tq (see Figure 4). The current will then rise to @ maximum level determined by the arc voltage, the maximum source voltage, and the source impedance. It is important to know not only the applied voltage waveform, but also the source impedance (which is equivalent to knowing the energy in the transient pulse) in order to find the peak current. Knowing the current through the gap is important because another characteristic of spark gaps is the capability of the gap to handle high peak currents. High peak current levels may physically damage a small gap by melting leads or burning away electrode material. For instance, a bigger gap that can handle higher peak currents may have a slower response time, and consequently ‘maximum peak currents must be considered along with other factors when the proper spark gap is being chosen.VOLTAGE Iv) w CURRENT Vasax 1 ' I 1 TIME (7) WAVEFORM WITHOUT SPARK GAP V SOURCE) ARC VOLTAGE Ta 8 WAVEFORM WITH SPARK GAP Figure 3 ~ Voltage Waveforms Zs0uRce Vsource IMAx = Vmax — Vane ZSOURCE Ta TIME (T) Figure 4 ~ Surge Current WaveformVOLTAGE (Vv) SARE SD SURGE VOLTAGE WAVEFORM WITHOUT SPARK GAP WAVEFORM OF SPARK GAP WITH SERIES RESISTOR (DISCHARGE VOLTAGE) WAVEFORM WITH SPARK GAP ONLY VaR + VaR TUME (7) Figure 5 — Surge Voltage for Gap with Current Limiting Resistor One way to alleviate the peak current problem is to attach a resistance in series with the gap. The series resistor has no effect until the gap breaks down and current begins to flow across the resistor-gap combination. (See Figures S and 6). The sum of the are voltage plus the voltage across the resistor, which in tum is dependent on the current, is called the discharge voltage. For 2 given transient source, the net effect of the series resistor Is to increase the discharge voltage while decreasing the peak current that flows across the gap. The voltage seen across the resistor-gap combination is greater than that which would occur across a gap alone, Thus somewhat less protection is offered by a protective device using a current limiting series resistor. In addition, any inductance of the resistor will tend to slow down the response time of the gap. PEAK CURRENT WITHOUT SERIES RESISTOR Imax = VAX — Varc ‘ZSOURCE PEAK CURRENT WITH SERIES RESISTOR CURRENT (1) lax = Vmax — vane ‘ZsouRcE +P tA TIME (T} Figure 6 ~ Surge Current Waveform for Gap with and without Series Resistor‘The next point to consider in basic spark gap operation is how the conduction of the gap is stopped ‘once an arc has been initiated, In terms of basic physics, the discharge will extinguish when the tlectron recombination rate exceeds the electron production rate, The gas will then proceed to become non-conducting and the discharge wil be extinguished. As with breakdown, this phenomenon is not very well understood theoretically, but an empirical relationship between voltage and current and “extinguishing” is shown in Figure 7. As illustrated on the volt-eurrent graph of Figure 7, a discharge will extinguish if the voltage and current are such that the appropriate point on the graph lies below the curve (see cross hatched area) while the discharge will continue if it lies above the curve, The current through the gap is primarily determined by. the source impedance. If the only voltage source is the transient against which one is protecting, the discharge will obviously extinguish 3 the voltage approaches zero, However, if the gap is protecting a DC system, the normal system DC voltage levels may be sufficient to keep the gap from extinguishing, For instance, ifthe power supply can provide enough current to keep the gap's operating point above the shaded area* of the curve shown in Figure 7 the gap will nt extinguish. This phenomenon is know as DC holdover. The curve in Figure 7 is a graph of the minimum DC holdover voltage as a function of current. Rather than siving the whole curve, spark gap specification sheets often give the DC holdover voltage at some given current (usually one ampere). The problem of DC holdover is that the gap continues to short-circuit the power supply with potential damage either to the supply or the gep itself Furthermore, cicuit operation cannot return to normal until the gap extinguishes GLOW RANGE Vea ee REGION| Vo = BREAKDOWN VOLTAGE % Vo = GLOW VOLTAGE 5 Ver. Vert. = EXTINGUISHING VOLTAGE 8 Va = ARC VOLTAGE s Va EE ELT] Figure 7 — DC Holdover Curve ee are eL * Its important to note that the volt-current behavior of a gap going into the arc is different from ‘one leaving the arc as indicated by the arrows on 10nner ovr In an AC system extinguishing the gap is somewhat different. Most gaps will extinguish because the sap voltage and current will return to zero every half cycle, The gap, however, may well conduct from the time the surge ends until the end of the half eyele in progress (See Figure 8). This AC circuit Phenomenon is called follow current. It occurs because the gap must continue to conduct during the remainder of the half cycle or until the voltage across the gap in combination with the current through the gap brings it into the extinguishing area. This follow current, even though limited in duration to a half cycle, can cause deterioration of the spark gap and result in changes of breakdown and insulation resistance, especially if the AC system ean supply large amounts of energy for the remaining half eycle A series resistor may be used to alleviate the DC holdover and follow current problems. As mentioned previously a series resistor will limit the current through the spark gap. In DC applications the resistor will tend to push the gap into the glow region (See Figure 7) where i extinguishes ata higher voltage level. Again, the disadvantage of the series resistor is that it lessens the protection against overvoltages 1.4 Types of Available Gas Tube Spark-Gap Protectors Gasttube spark gaps come in hundreds of sizes, many shapes, and have a wide variety of characteristics depending upon their intended application. A look at the data sheets and advertisements of the various manufacturers in electrical engineering and electronics publications fves one the impression that the choice is almost infinite. However most of the existing spark gap Protectors can be categorized by the following five general types or families ‘Twoelectrode, low-energy protectors ‘Two-lectrode, high-energy protectors ‘Three-electrode, low-energy protectors ‘Three-electrode, high energy protectors Hybrid Spark Gap protectors Each of these five types will be briefly discussed. Photographs and drawings of representative Protectors of each type along with examples of general specifications are included. The purpose of these examples is to indicate the wide diversity of spark gap protectors rather than provide a complete catalog. More information on specific devices can be obtained from the manufacturer's data sheets. Two-Electrode, Low-Energy, Protectors (0-20 kAmps) This type of spark gup consists of metal electrodes hermetically sealed to a glass or ceramic body to provide a controlled gas environment, high insulation resistance and low dielectric loss. The chamber between the electrodes is filled with @ gus or gas mixture. The gap may be obtained in a number of sizes with a variety of terminals. The Joslyn 2001 series MSP® spark gap isa typical example of this family of protector. The medium size unit in Figure 9 is this type of gap and is also shown in the cutaway drawing (Figure 10). The general specifications for this family of gaps are given in Table 2.TRANSIENT SUPERIMPOSED ON AC SIGNAL a (WITHOUT SPARK GAP) Vourace (Vv) AC SIGNAL AND 2 CLAMPED TRANSIENT ARC VOLTAGE VOLTAGE (Vv) GAP EXTINGUISHES SURGE CURRENT AND FOLLOW CURRENT Z CURRENT (1) FOLLOW CURRENT ONLY | periooor =I TIME (T) 14 SPARK GAP CURRENT Figure 8 ~ Follow CurrentFigure 9 — Photograph of Basic Two-Flectrode Low-Energy Gaps CERAMIC INSULATOR Figure 10 — Cross Section of Basic MSP GapTable 2 MsP © Spark Gap Specifications Voltage Ratings 230 V DC, 360 V DC, 470 V DC, 800 V DC, + 15% Discharge Current Rating 10 operations at 10,000 Amperes (10 x 20 ‘microsecond).* Repeated discharge 1000 mtd capacitor ‘charged to 750 V. Repeated discharge 300 Amperes (1 x 250 ‘microsecond).* Insulation Resistance Greater than 10 megohms at 160 V DC. Interelectrode capacitance <5 pf (Basic Gap) Temperature Range of Operation 100" F to 400" F Thermal Shock S cycles of heat and cold; each eycle coneists of 15 min, at 400° F Followed by 15 min, in ice water and back to 400° F for next eyele. Impulse Shock 1000 G on 6 + 1 millisecond wave, We 0.05 ounces Body Dimensions Nominal 5/16" dis. x 11/32" length. “See Appendix "B for waveform definition. ‘Two-Blectrode, High Energy Protectors (>20 kAmps) This family of spark gap protectors is very similar to the previous group except that these gaps are designed to handle very large currents without damage. The dividing line between low and high surge Protectors is not well defined, but certain differences in the design of high surge protectors make such a distinction useful Twoelectrode, high energy protectors are usually characterized by their large size and heavy, rugged electrode construction. The electrode construction is such that the energy from the arc willbe spread cover a larger area and has less chance to damage and erode the electrode material. Similarly, heat dissipation problems are less severe in a larger device. ‘The spinning arc technique is often used to help protect high energy devices from damage. In this technique the current flowing through the gap interacts with a magnetic field provided within the sap, causing the location of the arc to rotate, or spin, around the electrode. The arc motion is due to the same principle that causes the rotor of an electric motor to turn. However, ina spark gap of this type, the magnetic fields are created by magnets inserted inside the electrodes. The spinning arc technique is useful because it causes the are to continually contact new areas of the electrodes.Consequently, localized electrode heating and damage is minimized, and the life-time of the device is park gaps are most often used in protecting power lines where power fects may cause large amounts of current Twoelectrode high-energy poldove: flow through the follow and voltage protector. Another application of these high ene: 1s is for direct lightning stroke protectior where large currents are involved as specified in MIL-A.9094. Many such protectors are presently in use for HE radio protection on both military and civilian aireraft Figure 11 is a photograph of several two-electrode, high energy protectors manufactured by Joslyn The two devices on the left are used mainly in power systems protection while the protector on the right is an example of lightning protector used in the HF antenna system of many aircraft, Figure 12 is a cutaway drawing of one of these devices and specifications for a protector containing this device are given in Table 3, Figure 11 ~ Photograph of Two-Electrode High-Energy ProtectorsTable 3 Sp fications for High Energy Protector With integral Current Limiting Resistor System Voltage Rating (60 Hertz Sparkover Voltage DC Breakdown Impulse Sparkover Voltage Peak Surge Current Peak Amps Follow Current Max Discharge Voltage Operational Life Operating Temperature Range Pressure 120/208 V ems - 60 Hz 105 volts ems + 15% 250V + 15% 1200 volts at 10,000 V see 65 k Amps (10 x 20 4 5) 900 A peak half cycle 1000 volts (10 x 20 4 sec wave shape with 5 kA peak current 2000 operations at 10 kAmps 40° F 10 160" F Sea level to 12,000 ft Figure 12 — Cross Section of Joslyn 1903 Spark Gap HOUSING (METALLIC) 16 1 TERMINAL MSULATORThree-Flectrode, Low-Surge Protectors In addition to the previously discussed two-electrode protectors there also exist three-electrode, common chamber spark gaps. These three-electrode gaps have been found to be particularly useful in protecting grounded or ungrounded pair lines often found in circuit applications. The basic advantage of the three-electrode, single-chamber gap is that it eliminates much of the common mode voltage developed in a balanced pair line when separate, two-electrode gaps are used. A comparison between the three-clectrode gap and separate two-electrode gaps is shown in Figures 13 and 14. e 1 at the arcing voltage of t impulse breakdown In “A” of Figure 13 and 14 spark gap 1 ates fist at ¢y and clamps Lit spark gap 1. Because individual spark gaps ate unable to provide ex repeatability, and consequently are not capable of being matched for a particular application, spark ap 2 arcs at a later time ty depending on its impulse response to the rise rate of the input voltage transient During the time ¢2 - ty when Line 1 is clamped at a low voltage, the voltage on Line 2 continues to climb with the transient voltage rise ‘This transient voltage appearing across spark gap 2 during the time t2 - ty appears also across the load, less the low arcing voltage of spark gap 1. As a consequence the load to be protected is exposed to this common mode voltage for the time t -t]. LINE 1 LINE 1 \ ver / L1 ELECTRODE S863 L2 ELECTRODE vez LINE2 TWO SEPARATE TWO LINE 2 ONE THREE ELECTRODE ELECTRODE SPARK GAPS COMMON CHAMBER SPARK GAP ire 13 ~ Comparison of Two Separate Two-Electrode Gaps with One Three Electrode, Common-Chamber GapLINE TRANSIENT VOLTAGE WITHOUT GaP vr Tsar anes LI GROUND ARG | 1 | | | | | p> COMMON MODE VOLTAGE | — common move vouTace vio] | | | Time =] Time Tet we ‘SG2 ARCS 72 L2-GROUND ARC T2 TWO SINGLE CHAMBER SPARK GAPS ONE THREE ELECTRODE, SINGLE CHAMBER SPARK GAP A 8 Figure 14 — Development of Common Mode Voltage, Two Single Spark Gaps Versus One Three-Electrode, Spark Gap One method of protecting the Joad from this common mode voltage is to decrease or eliminate the time interval ta ty. The three-electrode, common chamber spark gap has this capability. In the common chamber spark gap, when an arc discharge begins from the Ly electrode to ground, its resultant gas ionization in the common chamber triggers an immediate are discharge to begin from the Ly electrode to ground, providing a voltage is present across the Ly electrode to ground. As a consequence the L2 electrode to ground discharge takes place very rapidly after the beginning of the Lj electrode to ground discharge, and ta - ty is reduced to a minimum. The fact that the formation of ‘one are creates ionization that tends to hasten the formation of a second are is known as sympathetic breakdown,As with the twoelectrode protectors, the three-electrode gaps are manufactured in a variety of shapes, sizes, and current handling capabilities. The three-eler low energy devices may often resemble two of their two-electrode counterparts simply connected in series. The difference is that @ common chamber is used and all electrodes share the same gas allowing sympathetic breakdown to A photograph of several representative low energy, threeelectrode, common-chamber protectors Is shown in Figure 15, Figure 16 is a cross section drawing of the same type of protector. Table 4 lists some typical characteristics of such a device Three-Blectrode, High Energy Protectors This class of protectors is primarily made up of larger versions of the three-electrode, common chamber spark gap already discussed, These devices are for very high current applications such as power follow or direct lightning stroke techniques such as the spinning are principle are used to help handle the large energies involved with sms. As with the two-electrode, high-surge prote: such protection devices cifications Representative three-clectrode hig fox h energy devices are shown in Figures 17 and 18 and spé typical device are given in Table 5, Figure 15 — Photograph of Low-Energy Three Electrode GapsTable 4 Typical Specification for a Low-Energy Common Chamber Three-Electrode Spark Gap Protector Line Application Voltage DC Breakdown Vottage Insulation Resistance Impulse Sparkover Voltage Mi 1um Holdover Voltage Capacitance Peak Surge Current Discharge Voltage at Peak Surge Current Balanced Operation Operational Lite Operating Temperature Range Shock Linear Acceleration CERAMIC INSULATOR 0-30 volts rms 400 Hz, 0-20 volts de 800 volts + 30% L-G 1000 megohms minimum at 600 volts de 1600 volts maximum at 40kV/s L-G 20 volts de Less than 2 picofarads 100 amperes on a 1x 10us waveshape 40 volts maximum Duration of metallie voltages above 80V less than 0.34 1000 operations with peak surge surge current and power follow 65° F to 350° F 40G for 11 millisecond, non-operating 106 for 1 minute operating ELECTRODE (GROUND) ELECTRODE Figure 16 — Cross Section of a Low-Energy Common Chamber Three Electrode Gap 20 RRA A CTE ET EEFigure 17 — Photograph of High-Energy Three-Electrode Gaps Pee Figure 18 — Cross Section of a High-Energy T Se IG (GROUND ELECTRODE) three-Electrode Gap a‘Table 5 Specifications - 40 coulomb three-electrode spark gap Protector Voltage Rating 300 V + 15% DC Sparkover Voltage 300 V+ 15% LG. 9300 V + 18% L-L Insulation Resistance > 10? ohms Impulse Sparkover Voltage 2.2 kV at 100kV/usee rise 5.0 kV at 1000kV/usee rise Minimum Holdover Voltage 30V Capacitance <50 pf Peak Surge Current, 7,600 Amps (40 coulombs total) Discharge Voltage at Peak Surge Current Operational Life 10 year shelf life S.consecutive 40 coulomb surges with no damage Operating Temperature Range 07°F 104257" F Pressure Sea level to 7000 ft. operating Shock 500 g's Hybrid Spark-Gap Protectors Thus far we have discussed protective devices that consist of spark gaps alone. Often, however, @ spark gap is used in conjunction with other circuit elements. The elements are often packaged together with the gap resulting ina class of hybrid protectors with certain advantageous properties. ‘The most common example of such a hybrid is the series spark gap — resistor combination. As discussed previously a series resistor can be useful in several ways. The resistor tends to limit the current through the gap, alleviating DC holdover and power follow problems. However, with a series resistor, the discharge voltage across the protector exceeds the arc voltage, due to the IR voltage drop ‘across the resistor and less transient protection for voltage sensitive cireuits is provided, ‘Various types of resistors are used in the construction of combination gap-resistance protectors. The simplest is the carbon resistor commonly used in many circuits, For higher current applications coils of heavy resistance wire, wound so as to minimize inductance, may be used. Another special technique is to use a protective device with a non-linear resistor. When a linear resistor device is used in high current applications the voltage drop across the resistor may become so large that litte protection is available (ie., the discharge voltage of the protector is high). However, special non-linear resistors exist that have volt-current characteristics which level off at a constant voltage for al large level currents. As a result, the discharge voltage of a protector using such a resistor becomes constant rather than increasing with current through the gap (See Figure $ in Section 1.3) and the protector clamps the line to a lower voltage than a lineat resistor device would. 22Other hybrid gap combinations are available. Two series gaps may be packaged together with the resulting protector having a glow voltage twice the glow voltage of a single gap thus assisting their ability to extinguish in a DC application. Several parallel gaps can be used to handle greater currents than single gaps alone. Inductors or capacitors are combined with gaps for special uses. Zener diodes are packaged with spark gaps to afford protection below the impulse response of the gap while providing the large surge current capability of the gap. Triggered spark gaps also exist where one electrode is designed to help initiate a discharge across the other two. In all it is readily seen that 2 wide assortment of combinations is possible. Examples of several types of hybrid protectors are shown in Figures 19 to 22. Specifications for several of these protectors are included in Tables 6 and 7. Table 6 Typical Specification for Low Surge Protectors Having High Series Resistance Protector Voltage Rating DC Sparkover Voltage Insulation Re Impulse Sparkover Voltage Minimum Holdover Voltage Capacitance Peak Surge Current Discharge Voltage at Peak Surge Current Operational Life Operating Temperature Range Pressure Shock near Acceleration Vibration Acoustic Noise Retistance to Oil (0-250 volts rms 400 Hz, 0-160 volts de £800 volts + 15% 1000 megohms minimum at 500 volts de 1600 volts maximum at 40kV/us 160 volts de Less than 1 picofarad S amperes on a 1 x 10us waveshape {600 volts maximum 1000 operations with peak surge current and power follow 65" F to 360" F 0.65 PSIA to 15.4 PSIA 40g for 11 millisecond, non-operating 10g for 1 minute operating 20:80 Hz Gdb/octave increase 120-200 Hz 1.4g# Jeyele/second 200-232 He Sub/octave decrease 2821000Hz 1g? cycle/second 1000-2000 Hz ——Sdboctave decrease 170db Mil-L-7808 Continuous Oil Mist Exposure 23Figure 19 ~ Photograph of Hybrid Protectors HOUSING SPARK GAP ORL RESISTOR TERMINAL Figure 20 ~ Cross Section of Typical Hybrid Protector an Table 7 Typical Specification For Multiple Packaged Low Surge Protectors Having High Series Resistance The unit consists of 2 separate protectors with a common ground Protector Voltage Rating 0-260 volts rms 400 Hz, 0.50 volts de OC Sparkover Voltage: 800 volts # 30% LG Insulation Resistance: 1000 megohms minimum at 600 volts de Impulse Sparkover Voltage 1600 volts maximum at 40kV/js LG Minimum Holdover Voltage: 50 volts de Capacitance: Less than 2 picofarads Peak Surge Current: 5 amperes on a 1 x 10s waveshape Discharge Voltage at Peak Surge Surge Current! 600 volts maximum Balanced Operation: Duration of Metallic Voltages above 8OV less than 0.3}. Operational Lite: 1000 operations with peak surge current and power follow Operating Temperature Range: -65" F to 350° F Pressure 0.65 PSIA to 15.4 PSIA Shock 409 for 11 millisecond, non-operating Linear Acceleration’ 10g for 1 minute operating Vibration 2080 Hz ——Edb/octave increase 80:200Hz 1.49" feycle/second 200-282Hz —_Sdb/octave decrease 232-1000 Hz 1g? /eycle/second 1000-2000 H —db/oetave decrease ‘Acoustic Noise: 170db Resistance to Oil: Mil-L-7808 Continuous Oil Mist Exposure HOUSING oat? DS SPARK GAP RESISTOR SS) yma RoR SSS a Figure 21 ~ Two Electrode Gap with Series Resistor 25HOUSING X Noe ae SSN. Wenig Figure 22 — Three-Electrode Gap with Series Resistors THREE-ELECTRODE GaP: LEADS GROUND ELECTRODE 1,5 Performance Information ‘There is no available performance information on the ability of a spark gap, or any other protective device to withstand an actual nuclear environment. However, spark gaps have been tested to Simulated EMP threat levels with varying results. All the Joslyn devices used for EMP applications have successfully met the test specifications provided by the customer and field reports seem to verify the calculations and predictions made for their life and operation. Usually these devices are used for Protection from non-nuclear transients and most of the performance information is based on non EMP applications (Reference 3). For this book the representative families of spark gap protectors described In Section 1.4 were tested for EMP voltage rates of rise. The following is a description of ‘the test procedures used for these tests, The results are shown in figures 25 through 29, In order to obtain the response test data shown in Figures 2SA through 29C two basic test circuits were used, Figure 23 shows the cicult digram for slow to medium rise time pulses, yielding rates of rise between 10Vusee and 100kV/us. The pulser was constructed in such a way that the charging voltage (Eo) could be varied and the storage capacitor (C), the series inductor (L), and the current limiting resistor (Rser.) could be easily changed by means of plug-in components. Figure 24 shows the circuit diagram for fast rise pulses above 100kV/sus and up to 2kV/nsec with a Peak voltage of 20kV. Rates of rise were varied by changing the charging voltage on the coaxial storage line The voltctime curves in Figures 25A-29C show typical response characteristics. All devices were tested at least 3 times in both polarities for each major rate of rise. Lead lengths of test devices were kept to 4 minimum especially forthe fast rise measurements. The data for fast rise impulse testing of Joslyn PIN 1919-04 (Figure 26C) was obtained from the Lightning Transients Research Institute, Minneapolis, Minnesota,1MQ. t RSER. TEKTRONIX 507 OSCILLOSCOPE ZIN = 520 Figure 23 ~ Slow Rise Impulse Test Circuit ‘s52db cHARGI ate SPARK ATTENUATOR RESISTOR GAP TEKTRONIX 519 OSCILLOSCOPE STORAGE LINE HY, SUPPLY Figure 24 — Fast Rise Impulse Test Circuit, aDIMPULSE BREAKDOWN (VOLTS) IMPULSE BREAKDOWN (VOLTS) 10° 10° 10 10 10 10 Figure 25 ~ Volt-Time Curves for Two-Electrode, Low-Energy Protectors | e|| s AA | GS Figure 254 — MSP® P/N 2001-02 wr os 28 10° 10° 10" TIME (SECONDS) igure 25B ~ Small MSP® P/N 2301-14 10* 10° soovius 420-680 515 14kVisis 510-690 585 TORvius 640850 739 YOOKV/us 850-1250 ro32 LARVins—2kV@s.2kVInt 1900 qe Rowe Range of Average of Rise Breakdown Breakdon oc 792.860 837 Vins 800-1030 ‘802 rkvise 8201070 956 YOKVin 10301300 149 YooKVis 1350-1600 1483 13kvis = 250.9 3000 BOKV © 1.1kVineIMPULSE BREAKDOWN (VOLTS) IMPULSE BREAKDOWN (VOLTS) 10° 10 10 10 10 of Rise oc rooviis a tvs revise SH rooKVius 1.5kVins wea rime (seconos! Figure 25C — Jumbo MSP® P/N 71506, ‘| 3 ( x Rate s ose g|is é oc < RS VoOV/us | $ hakvis tevin | Migs lv tine oss? me (seconos) Figure 25D — HD Jumbo MSP® P/N 71513-02 Range of Breakdown 327621 490770 590.70 760-1030 1101480, 22kV @ 9kVins 21618 am.ag2 430670 510-690 730.920 1050-1350 29 432 28 028 227 94 25KV@ 1.4kV/ns 2200 180.6IMPULSE BREAKDOWN (VOLTS) IMPULSE BREAKDOWN (VOLTS) 10° 10 10 10° 10 10 Figure 26 ~ Volt-Time Curves for Two-Electrode, High-Energy Protectors ie = g A ¢ ’ s 8 A | Md esp20h SAG Rd PIN jena : ¥ XT 4 = s MU] Le | le $. cI Poe oor SE Gea resets 30 ‘ev ae asc ae Raw ange of oF Rise Breakdown oc 507-516 509 woovins 60570 565 vevins 720750 740 1OKViEs 10-1190 ns7 1OoKVIe 2950-2550 2458 12kVins— 5.2kV@1.9kVins 4600 a9e.9 Raw Range of oc 432.443 430 voovins 40.810 20 tkvipe 700.940 807 wORVis 10201940 7 WOoKViis 17002200 1933 VARVIng 44k @ 1.7KVins 300.7 3700yours) IMPULSE BREAKDOWN IMPULSE BREAKDOWN (VOLTS) w S ot w 10° Figure 27 ~ Volt. 10* 10° 10+ 10% 10" 107 ‘TE (SECONDS) Figure 26C ~ Aircraft Lightning Arrester P/N 1919-04 ime Curves for Three Electrode, Low-Energy Protectors 10 0 10 I 1 4 x NC ny & ° a rime (seconos) Figure 27A —Trigard® Lineto-Line PN 71508-04 (TABULAR DATA WAS NOT AVAILABLE) Rate Average of Rise Breakdown oc 360 75viEs 600 veins 660-1320 894 TOKV/He 1050-1810 1201 tookVius 2000-2400 2210 B1e22 31IMPULSE BREAKDOWN (VOL TS) IMPULSE BREAKDOWN ivoLTS) 10 10 10 0 10 TIME (SECONDS) Figure 27C — Balanced Trigard® Line-to-Ground P/N 2021-35 32 s ae st | Pd mee tons igs gos we Seconos) Figure 27B — Trigard® Line-to-Ground P/N 71508-04 s ¢ st fll sl le ells NAL f roe rose rose ere wa Rate Range of Avera, of Rise Breakdown Break, be 170-190 180 Yovins 290.980 470 tkvius 43061100 652 TOKV/Es 6901190 885 7SKVis 1100-1700 1419 TOokvius 1450-1800 1625 V9KVins —3BKV@TkV/ne 5200 78019 Rate Range of Average of Rice Breakdown Breakso be 720-962 880 FERVins 8201120 951 tkvius 9001250 1100 YORvius 12801600 1420 WO0KVius 1500-2000 1747 LAKVing — 2.0KV@ BkVins 3500 35kV 024IMPULSE BREAKDOWN (VOLTS) PULSE BREAKDOWN (VOLTS) 10° 220} ae 4d 10 tw TIME (SECONDS) Figure 27D — Balanced Trigard® Line-to-Ground P/N 2021-35 ¥ 10" se a] s a1 | AH A HW 10! 10! mew? wos wt TIME (SECONDS) Figure 27£ — Two Coulomb Balanced Trigard® Line-to-Ground P/N F5003-01 average Breakdown 29819 461 voovius 500-770 572 1akvins 540-890 623 toevius 670-950, 835.4 TooKvis 950-1650 v2 LAKVIng SV @ 4kVIne 2200 26812 (TABULAR DATA WAS NOT AVAILABLE)34 IMPULSE BREAKDOWN (VOLTS) IMPULSE BREAKDOWN (VOLTS) Figure 28 — Volt-Time Curves for Tree Electrode High-Energy rotectos ¢ 10! = x § é 3 g} | ei} ls lan FF |1¢ 10 10! nrg eee o eomdtet tebe teste ven TIME (SEcONOS) Figure 28A — 40 Coulomb Trigard® P/N 5007-01 w pot Aa & § s ¢ & | fa KIS | x a 10! a a TIME (SECONDS) Figure 28B — HD Trigard® Line-to-Ground P/N 2003-10 10? (rasucan DATA WAS Nor AVAILABLE) (rasuLaR DATA was Nor AVAILABLE)x ot < - N tT (TABULAR g DATA Was 3 s A Nor = = AVAILABLE) 8 p se fo é 4 3 ~ rAd = 10 wt? Te (sécoNos) Figure 28C — HD Trigard® Line-to-Line P/N 2003-10 Figure 29 — Volt-Time Curves for Hybrid Protectors re + ly a x * ate Range of 5 orice eas g f\ ls 2 é oe o7eaan 2385 8 < xt oovias 450.770 s7 8 sts g = $ tkvins 810740 x3 z aed sokVins 800-1020 872 4 ooKviis 170018501386 2 “] —tkvies_ 38003700 3600 = 1.6kV/ns_— 4500-4700 4550 0 Lo a a TIME (SECONDS) Figure 29A — AC Arrester P/N 2301-38,IMPULSE BREAKDOWN (votTS) IMPULSE BREAKDOWN (voLTS) 10° 10 10 10 0 < é ‘il, rt TE (SECONDS) Figure 298 — AC-DC Protector P/N F5011-01 Vay Zt £ iar, 36 10* 10° 10+ 10 10* 0? ME (SECONDS) Figure 29C — AC Arrester P/N 2301-08. woovius —-s20680 tkvis 650.850 TOrVins 9801180 7kVIs 1400-1700 Yo0KVius 1080-1820 NASKVIns—35:4.2kV Brown oc 391-420 r00vlus 410540 tkvis 70680 Skvius 660-750 ToWvius 9001050 YOOKVins 1420-1750 thVies 31003600 Averaye Breaks 396 402 620 707 98 1595 4002.0 EMP Considerations 2.1 Background ‘The EMP or “radio flash” effect is the name applied to an electromagnetic disturbance generated by a nuclear detonation, however, EMP effects have been observed in high explosive as well as nuclear detonations. Theoretical physicists predicted EMP during the development phases of nuclear ‘weapons. EMP protection was provided for test instrumentation and a number of EMP measurements were made during the various nuclear tests. During the Fish-bowl series of tests in 1963, effects were recorded in widely separated locations indicating that EMP from high altitude detonations could pose serious problems for unprotected communications and power networks. Star Fish, 2 1.4 megaton detonation at 400 kilometers altitude, caused a number of outages and other disturbances as far away as Honolulu (see Reference 4). ‘The primary interest at that time, however, was focussed on other nuclear effects, such as blast, thermal, and electromagnetic blackout. Although EMP was recognized as a possible threat it was not widely understood. ‘After many years of research involving the improvement of EMP theory, the analysis of nuclear test data, the design of large computer codes to calculate EMP effects, and the development of EMP simulators which can demonstrate the effects of EMP, the seriousness of the problem has been generally recognized. ‘Asa result a number of military defense systems which must operate in a nuclear environment have been hardened against the effects of EMP. EMP protection criteria and specifications are being included in design and test requirements for new systems, and existing systems are being assessed and upgraded as needed. These systems include aircraft, missiles, satellites, a wide variety of ground based communication systems, and some power systems. ‘The details of EMP phenomenology and effects are largely documented in a number of classified handbooks and source documents which cannot be cited here. These contain comprehensive and up to date information on the subject and should be consulted if possible, Recently « document (ceference 5) was published in the open literature by the Defense Nuclear Agency which is much ‘more thorough than the simple treatment we have provided here and is recommended to the reader for further study of EMP considerations. 372.2 EMP Environments General ‘he electromagnetic pulse following 4 nuclear detonation produces intense transient electric and magnetic felds with very short rise times and a frequency spectrum extending from almost sere to ‘more than 100 MHz AAn analogy to EMP from ordinary experiences is the electromagnetic radiation from lightning. However there are large differences between the characteristics of the two phenomena, A detailed tiscusson of lightning characteristics and protection techniques is given in a companion dock Fectrical Protection Guide for LandBased Radio Facilities, by David Bodle (Reference 9. Information on the difference between EMP and lightning is given in References 5,7 and 8. in this section we have chosen to use the tile EMP Environments because the electromagnetic pulse Produced by a nuclear detonation can have different characteristics depending upon the alitude of the detonation, the location of the system in relation tothe burst point andthe design of the nuclese weapon, Basically, the environments can be divided into 3 types — ground-burst EMD, aibuist EMP and ehaltiuds EMP. Ground-burst EMP includes bursts on or neat the surface, ie. keine prproximately 0-3km, air-burst EMP includes bursts from between approximately 0.3 and 40k; and high altitude covers bursts above about 40 km. [isu 30 depicts the diferent types of bursts and the area of intense electromagnetic felis produced by each. ste TMS mon acrrune nunsr HIGH ALTITUDE DEPOSITION REGION AIR BURST GROUND BURST 4 4N Figure 30 — Simple Comparison of EMP Regions 38Bursts on the ground and within the atmosphere produce intense fields in the source region, or deposition region, but this region covers a relatively small area. High-altitude bursts, on the other hand, can produce significant fields over wide geographical areas and probably pose the greatest EMP threat to susceptible systems. It should be noted that all bursts radiate signals which ean be detected at great distances, Generation Mechanisms ‘The primary source of EMP is gamma radiation. Gamma rays are high-energy photons, emitted directly from the fission process in an exploding nuclear bomb. These photons are also emitted by the interaction of neutrons from the bomb with the air or ground near the burst point. About 0.1% of the energy of a typical nuclear bomb appears as prompt gamma rays. As the stream of high-energy photons moves through the air Compton* scattering collisions occur and a current of approximately ‘one-MeV electrons is produced. It is this electron current which is the direct cause of EMP. ‘The Compton scattering phenomenon, depicted in Figure 31, is a process in which a gamma ray collides with an electron (e.g, in an oxygen or nitrogen atom) and propels it in or near the forward direction of the original gamma ray. This process is repeated for each of the prompt gamma rays in a very short period of time. The large number of outward-steaming Compton recoil electrons constitutes an electric current, called the Compton current. This current gives rise to electric and ‘magnetic fields, SCATTERED GAMMA RAY INCIDENT GAMMA RAY RECOIL ELECTRON Figure 31 ~ The Compton Effect Compton electrons, in passing through the air, free large numbers of low-energy secondary electrons from atoms by ionizing (or Coulomb) collisions. These low-energy electrons strongly enhance the electrical conductivity of the ait. Thus the electric field, generated by the Compton current, in turn produces a conduction current in the air. The conduction current generally tends to eventually cancel the Compton current, and to limit the EMP generated. * In 1924 a physicist, A.H, Compton, discovered that the collision of photons with electrons can dislodge the electrons from their atomic bonds, 39Ifa spherically symmetiic bomb were exploded in a uniform atmosphere, only a radial electric field would be generated, which would exist only in the vicinity of the burst point, and electromagnetic waves would not be radiated large distances. Realistically, however, asymmetries in the bomb and its environment lead to the production of dipole electric and magnetic fields and to electromagnetic waves which propagate large distances. The chief environmental asymmetries are the air-ground interface for low altitude bursts no more than a few kilometers above the ground, and the air density gradient and the earth’s geomagnetic field for bursts at higher altitudes. For burst altitudes above 40 km the earth's geomagnetic field creates a very strong ‘asymmetrical force on the Compton electrons. In the reduced density of the atmosphere at around 30, km this force deflects the radial Compton electrons, and therefore current, into transverse directions, and the transverse currents radiate as if propagated from a well-designed antenna. Although gamma radiation isthe primary source of EMP generation it shoud also be mentioned for completeness that X-rays may also give rise to EMP thcough ther photoelectron production, Even the intense plasma bubble created in air by a nuclear detonation may be responsible for generating EMP by temporarily excluding the earth’s magnetic field from the burst region, Ground-Burst EMP. When nuclear detonation occurs at the earth's surface, gamma rays and the resulting Compton current flow approximately radially from the burst point. The initial result of the radial Compton current isa radial electric field caused by the separation of positive and negative charges. In the air, at distances less than two kilometers, this radial electric field builds up until it drives a conduction current which balances the Compton current, At these short distances the radial electric field saturates and increases no further. The peak value of the saturated radial electric field is thousands of volts/meter with a very short rise time. ‘The gamma rays and the Compton electrons penetrate only a few feet into the earth. However, due to the relatively high conductivity of the earth, the radial electric field in the air just above the earth is disturbed and the conduction current near the earth tends to flow in the ground. This sets up current loops, with Compton electrons flowing outward from the detonation in the air and conduction currents flowing forward to the burst point in the ground. These current loops produce ‘an eximuthal magnetic field, strongest at the Earth's surface but diffused by the skin effect into the air and the ground. The peak value of the magnetic fied can be on the order of 100 gauss with a very short rise time. Both fields, the radial electric field and the szimuthal magnetic field, decay in a short time, on the order of microseconds. In the air, the vertical and transverse dipole electric fields decay much more slowly and maintain values of many volts per meter for a long time. However, the power that can be delivered by these fields is limited by the increasing resistivity of the air caused by the attachment of electrons and the recombination of ions, 40‘The field descriptions given thus far apply to the source region (deposition region) within a few kilometers from the burst, where the Compton current and conductivity are important. At greater distances the later effects are neglble; the electric and magnetic fields at these distances propagate as waves generated in the source region and fall off as the reciprocal of the distance. This is due to the ar-ground interface which allows only upward Compton curtent, and produces an electric dipole signal for distant observers, ‘Air-Burst EMP At higher burst altitudes, there is a decrease in the azimuthal magnetic field, the transverse electric field and long distance field radiation. The radial electric field does not depend upon the ground for its existence. As the burst point is moved upwards, there is at first an increase in the radius of the region inside which an intense radial electric field is produced due to a decrease in attenuation of the gamma rays by the air. This holds to altitudes at which there is not enough air to make a strong Compton or conduction current, whereupon the radius of the intense electric field region decreases, Air bursts produce some radiated fields due to the earth’s geomagnetic field and air density gradient asymmetsies, but they are generally smaller than for ground bursts and much smaller than for the radiated fields from high-altitude bursts. High-Altitude EMP ‘The gamma rays from high-altitude bursts travel outward in all directions. Because of the extremely low air density, there is little attenuation of the horizontal and upward rays. However, the gamma rays that travel downward toward the earth eventually encounter the denser atmosphere and produce ‘Compton recoil electrons. At an altitude of approximately 30 km the gamma ray attenuation length becomes equal to the atmospheric scale height. This is the source (deposition) region for high altitude EMP. ‘The Compton cusrent from a high altitude burst is intially directed downward, but is soon deflected by the earth's geomagnetic field and 2 transverse current results. The transverse current generates a radiated EMP which continues to propagate downward radially from the burst point and is also partially reflected by the ground. Objects in and below the source region are exposed to this pulse, ‘The EMP from high-altitude bursts covers large geographical areas. An area of coverage of 1200 miles radius for a detonation at 300 kilometers is cited in Reference 5. Although the actual EMP threats are classified, some of the basic characteristics of a representative EMP from 2 high-altitude burst have been published in unclassified literature. Reference 9 gives the following: ““A representative electromagnetic pulse from a high-altitude bust will typically have maximum field strengths near the ground on the order of SOKV/meter, time duration on the order of a microsecond and rise times on the order of 10 nanoseconds, resulting in broad frequency effects to systems and equipments and dampened exponential ringing of circuits at their fundamental and harmonic frequencies.” 41More information, inchiding plots of electric field time wave forms and calculations of representative induced transients, can be found in References 10 and 11 2.3 Coupling Modes System Definitions Unfortunately for the engineer concemed with designing protection for critical circuit elements, an understanding of the EMP environment is only part of the knowledge needed. The transient current or voltage waveform seen by a specifi transistor or integrated circuit may bear little resemblence to the shape of the incident EMP waveform discussed previously. This is due to the fact that the ‘espons: ofthe overall system determines the coupling between the incident EMP and some particular circuit element. ‘an example of system effects, considers missile in flight. Assume the mis is illuminated by the EMP from a high-altitude burst sufficiently far away so that blast and radiation effects con be ignored. If the outer skin of the missile isa perfect conductor with no holes or penetrations it will seve @$ a Faraday cage and no transient will penetrate to circuits within the missle, The incident field, however, would induce currents on the skin of the missile. A long, thin missile would look ‘much like a dipole antenna and thus the skin currents would tend to flow back and forth on the ‘isile at some resonant frequency determined by its length. These skin currents eventually readiate the induced energy so that the skin current waveform resembles a damped sinusoid. Note that both the frequency and decay time of such a damped sinusoid are functions of the shape of the missile {ather than that of the incident EMP waveform. The magnitude ofthe incident field in the frequency ‘and of interest is important, however, in determining the magnitude of the induced skin current Now consider a more realistic missile with numerous aperatures, antenna feedthroughs, hatches, and other deliberate penetrations in its metal skin. The skin current will resonate about the same as it would without the penetrations, but all the “holes” in the metal shield will allow various amounts of nergy to couple into the missile. The fields that penetrate the skin often have waveforms characterized by the dominant resonance of the induced skin currents. However, the waveform may. well be changed again dve to the nature of the penetration. A cavity within the missile may be excited and ring at its primary resonant frequency, or an actual antenna may change the shape of the waveform by responding to certain frequencies better than to others. By the time the transient reaches some critical circuit component, its waveform and energy content will be considerably different from that of the incident EMP. Hence, the design engineer is forced to think in terms of overall system response to EMP rather than considering simply a specific incident EMP signal driving the particular circuit of interest. The overall System response can be very complicated. Theoretical calculations are extremely difficult due to ‘System complexity, and experimental tests are often extremely expensive, if possible at all, on large ‘weapon systems. However, some simple means for characterizing types of systems and ther texponses ‘can be very useful in understanding the overall problem, 42 LLL TL LEE LCOne possible way to characterize a weapon system is to use the system's location and expected ‘operational environment, From these the design engineer can postulate the nuclear effects the system, may be required to withstand. Satellite Systems The primary EMP threat to a satellite ean come from either of two sources. The first is the radiated EMP from a distant burs. If this signal travels through the ionosphere, dispersion effects will greatly change the shape of the incident waveform, eliminating the lower frequencies and dispersing the higher frequencies. A second threat is due to what is known as system-generated or internal EMP. If a nearby nuclear burst illuminates the satellite with numerous x-rays and gamma rays, the photons produce Compton electrons and photoelectrons, which in tum produce large ‘loctromagnetic fields. Missile Systems ‘The EMP threat to a missle depends greatly on the location of the missile. A missile underground in a hardened silo is obviously a different problem than that of a warhead approaching. its target. In its silo, the missile system includes all the associated ground equipment, power supplies, communication links, and control equipment, Nearby ground bursts may be expected, hence planning must take into account blast, shock, and radiation effects as well as EMP. After launch and during powered flight, the ground equipment is no longer part of the system and the shielding effects of the silo are no longer present. After engine burnout, the delicate electronics of the guidance computer may no longer be of importance, but the warhead fusing mechanism obviously is of prime importance. The characteristics, and consequent protection problems, of such a system change as the mission progresses. Aircraft Systems One may question the need to protect aircraft from EMP effects since many aircraft have flown in the vicinity of nuclear tests without apparent damage. Nevertheless, a potential EMP problem exists due to the increasing use of sophisticated electronics in aircraft, including computer directed navigation and weapon delivery systems. The prime EMP threat to aircraft is the high-altitude burst, simply because at the limited range of EMP effects at low altitude the other ruclear effects become much more important. Ground Systems ‘The term “ground system” is very broad and covers a diversity of military hardware including ships, tanks, large radars, satelite tracking stations, etc. Any of these systems may contain electronic equipment requiring protection from EMP, Radars are designed to be extremely sensitive receivers of electromagnetic radiation. Telephone switching centers, highly important for effective command and control, are connected to long lines which tend to collect and focus EMP energy. Computer facilities may be upeet by transients induced in power lines. In all, it seems that a great number of circuits need protection against EMP-induced transients ‘The main reason for briefly reviewing these various categories of military systems is to point out the vast diversity of devices requiring EMP protection and to note that the problems of each of these systems may be completely different. An unshielded radar obviously has problems different from those of a missle sitting in a concrete and steel enclosed silo. A designer working on EMP protection ‘must consider the overall system with which he is working. 43External Coupling AS in the example of a missile in flight, it is sometimes convenient to divide the EMP coupling Problem into 2 series of steps proceeding from the incident signal toward the particular circuit oF component of interest, The frst step in this sequence is frequently called “external coupling”, External coupling may be viewed as the overall response of the system to the incident fields. In the ‘missile example, extemal coupling would refer to the skin current and charge density induced on the conducting skin of the missile by the incident EMP. For an aircraft, information on the induced current in wings, Fuselage, tail and other exterior surfaces would be the goal of extemal coupling analysis or experiments. For a radar, external coupling analysis considers the effect of the incident EMP signal on the antenna. The concept of external coupling is to attempt to simplify the problem by finding some overall response of the system, This overall response is usually characterized by certain gross electromagnetic features of the system such as the dimensions and shapes of the principle conducting surfaces; e.g., the skin of a missile. In many cases these conducting surfaces form at least a partial shield around various electronic packages and are thus the external parts ‘of the sostem that first “see” the EMP. It is also known experimentally that waveforms measured deep inside a system may have dominant frequencies that are determined by the prime resonances of the extemal structure, Thus an understanding of external coupling is necessary in determining what goes oon inside. One should note that itis sometimes difficult to decide what is the “extemal” part ofa system. For example, a missile or aircraft may have fiberglass rather than a metal as its skin. In such eases, the interior electronics are illuminated directly by the incident EMP because the external surface offers essentially no shielding. Internal Coupling ‘The next step in this coupling sequence is labeled “internal coupling.” Here the analyst is concerned with how the transient energy gets from the outside to the inside of the system because no metal skin or shield is ever really continuous. Small holes, faulty RF seals, windows, and deliberate penetrations for signal cables, power lines, water pipes, etc., always are present and allow some transient fields to leak into the system. Once the energy enters, coupling take place in a variety of ways. On an aircralt, the incident EMP fields may leak through the doors of the bomb bay into a cavity (tho bomb bay) that resonates at the cavity-size-dependent frequency. Hundreds of cables may run through the bomb bay and cunrents induced on these cables may be carried throughout the aircraft to critical components. The coupling analysis is thus complicated because of the complexity of the system, and standard circuit analysis techniques are difficult to apply since the various fields form a distributed source which drives the entire circuit all at once. 2.4 Hardening Techniques Tracing the path of the EMP signal finally leads to some critical circuit element where the induced current oF voltage waveform will cause transient upset or even actual physical damage to the component, For example, the EMP induced transient may reset a flip-flop of a digital circuit in a 44 LLDMBMAAOPPVA PLP PLL LLL,computer or control device resulting in @ missile missing its target. Enough energy might be deposited in the junction of a transistor or integrated circuit to cause bum out or changes in operating characteristics, Another possibility is the inadvertent shorting out of a power supply due to arcing and insulation breakdown in its connectors and cables. Thus, it can readily be seen that a large number of devices may require some EMP protection. Although iti vitually impossible to eleminate all the transients induced by an electromagnetic pulse, certain general design guidelines can be used to greatly reduce the amount of energy reaching the ‘most critical circuit elements. These guidelines can be used to “harden” the system to EMP effects so that there will be less chance of logic upset or component damage Because these hardening techniques are highly dependent on the properties of the overall system, they are best applied when the system is first designed. Hardening a soft system can be very expensive. In this section we discuss certain general principles including circuit layout, shielding, ‘grounding, and cabling. Specific protective devices are covered in later sections. Shielding Deliberate electromagnetic shielding is probably the best way to harden a military system against EMP effects. Screen rooms can be built to attenuate fields by several hundred db, but military systems cannot always be placed inside shielded enclosures. Aircraft must have doors, windows, and 2 myriad of ports and hatches as well as external antennas. Also, the airframe weight must be as low as possible for maximum payload and performance. Thus it is readily apparent that in real life conflicting requirements often result in compromises in any electromagnetic shielding system. Many analytical and experimental studies have been directed at understanding shielding and the penetration of electromagnetic fields through imperfect shields. Here we discuss only the general idea of what makes shielding imperfect; ie., how energy penetrates practical shields and how these penetrations are minimized. Pethaps the most obvious fact is that no real shield is perfect. Numerous “holes” or penetrations, both deliberate and accidental will exist in any metal shield. Joints in metal skins are often built for mechanical integrity rather than electrical continuity, access doors or windows may not be fully closed, and RF gaskets may become wom or faulty. Rubber seals may resist gas pressure but pass EM radiation, Fiberglass or other non-conductors may be used for surface panels in situations where the electrical effects are considered less important than other factors. Or the skin may be simply so thin that it offes little protection against low-feequency signals (particularly the magnetic field). Deliberate penetrations include a variety of antenna feed-throughs, communication lines, or power distribution systems ~ all designed to couple specific types of energies into the shielded enclosure. An example of the problems posed by such penetrations isthe out-of-band response of an antenna. Often not even considered by the antenna designer, this characteristic may provide transparency in the frequency regime of the prime energy band of the incident EMP and thus may result in the direct coupling of damaging energy to delicate receiver electronics 45Mechanical considerations provide another class of deliberate penetrations, for example water pipes and sewer lines. To EMP, such penetrations may well appear as very good antennas leading directly through the shield. The designer must consider the electrical implications of each such penetration as carefully as he considers mechanical efficiency. ‘A conducting cage with a minimal number of penetrations surrounding the system being protected van provide adequate shielding. This shielding can be added specifically to protect against EMP, ot the «existing structure with a few modifications can be used as a shield wall. Multiple shields, one inside the other, can be used with great effectiveness. However, no shield is really “perfect” and the engineer should always be alert in order to detect and eliminate as many unnecessary penetrations as possible. Cabling Since long lengths of cable will act as efficient antennas for receiving EMP, itis fairly obvious that the uwse of properly designed and shielded cables will add protection to a system. Shielded cables are better than unshielded, tightly braided shields are preferable to loosely woven ones, and solid shields are the best of all. Unfortunately, cable costs may rise dramatically as the shielding quality increases, ‘Some cables require shielding for reasons other than EMP protection. Cables carrying high frequency Signals are often shielded to prevent pickup of RF noise from other sources. Such cables are probably “hard” to EMP damage. On the other hand a vast number of power lines and low frequency ‘communication links use cables with little or no shielding and all such cables should be considered antennas when considering the EMP response of a system. Also, a typical problem with shielded cables is the use of improper connections at the cable ends. If they are not fastened properly initially, or work loose with use, they may allow energy to couple to the inner conductors, rather than harmlessly draining through the shield. In order to make the use of shielded cables worthwhile, one must take cere to prevent such leakage, Grounding ‘Traditional ideas about grounding must be carefully considered when applied as techniques for ‘hardening systems to EMP damage. For instance, the concept of driving a metal stake into the earth and running a connecting wire from the stake to the grounded end of a circuit may be of no value at all with respect to EMP protection. At the high frequencies seen in an EMP, the connection between the stake and the earth is @ high impedance. In addition, the connecting ground wire may actually ‘couple energy to the circuit rather than away from it, More complicated grounding schemes may be useful, but even well-designed grounding systems for lightning protection may be useless for EMP purposes because the EMP signal contains higher frequency components. In addition, differences in the electrical parameters of different kinds of soil ‘make general purpose grounding schemes difficult. Internal grounds that provide a common reference potential for a variety of circuits are a somewhat we P 46different story. Such grounds may be somewhat useful if the connecting wires do not form ground loops or act as receiving antennas for the EMP. Thus it can be seen that traditional grounding methods must be carefully examined for EMP protection as they cant even serve to pick up more EMP energy if improperly used. Circuit Layout Careful consideration of the details of circuit layout may aid in minimizing EMP coupling. At the circuit board level some attempt is usually made to minimize inductive and capacitive coupling due to RP interference problems between circuits. The same care must be taken on a somewhat larger basis, to minimize EMP coupling. One must always remember that long wires act as elects dipole antennas and that loops are magnetic dipole antennas. Thus citcuits should be atranged so as to have the shortest possible connecting wires, and inductive loops should be avoided. The circuit designer should also remember that the EMP signal is broad band, and includes a significant high frequency content. ‘The distributed inductance and capacitance of various circuit elements must be considered. If there is @ choice of circuit location, the shielding effects of conducting walls or surfaces should be ‘used, In general, the further from an aperture the less the field strength. Care must be taken however, not to locate circuits near inadvertent antennas that may couple large amounts of energy into the systems. Examples of such inadvertent antennas include cable bundles and power lines that penetrate the metallic shielding. In addition, such non-electrical systems as water and sewer pipes and hydraulic lines must be avoided since such systems can carry large currents which may couple to more critical electronic systems. ‘Thus the basic point to remember is that every conductor in a military system is part of the overall circuit as far as EMP is concemed. The effect of all these conductors must be considered when the layout and location of critical circuits are chosen. Protection Devices In tracing the EMP energy from its source (a nuclear burst) to some critical circuit element it was noted that in the last step of the coupling process the EMP is characterized at a given point by some voltage transient with 2 given source impedance (i.e. energy content). The system can be hardened by introducing a protection device to eliminate or at least change the nature of this waveform before it reaches critical circuit elements. These protection devices are a class of circuit elements designed to reduce the effects of transients on the normal operation of the circuit. Included under the category of protective devices are: various filters and chokes, fuses and circuit breakers, silicon controlled rectifiers (SCR’s), zener and other diodes, spark gaps, gas tubes, transformers, non-linear resistors, as well as many hybrid designs combining several of these concepts. It is readily seen that there is a wide variety of protection devices. A description of the characteristics of each of these is beyond the scope of this document. (References 5, 9, and 10 give more detailed descriptions of the various devices mentioned above.) It should be noted that each of these devices 47