ML20125C048
| ML20125C048 | |
| Person / Time | |
|---|---|
| Site: | Zimmer |
| Issue date: | 11/13/1979 |
| From: | Boettger J, Haralampu G, Koepfinger J POWER ENGINEERING SOCIETY |
| To: | NRC OFFICE OF THE SECRETARY (SECY) |
| Shared Package | |
| ML20125C050 | List: |
| References | |
| RTR-REGGD-01.XXX, RTR-REGGD-1.XXX, TASK-OS, TASK-RS-705-4 NUDOCS 8001030106 | |
| Download: ML20125C048 (278) | |
Text
{{#Wiki_filter:O' ,
- 4. GWGMITB9NG gcc]rfy ,
o, H * ,, Y. t%, ,. .-o.%idr. s s X :. im
. l- ,m
~ w o y e-
- 7. gg7pOlardltmpt-W Englan
. Re h g,dM*fve& J.'.actr c/ System
. < ,orough, MA ) 8:
CO > l November 13, 1979 l N """ M' c., OchCc l PF0 POSED rut.E ] , gd6 Secretary of the Commission Docketing and Service Branch. l U.S. Nuclear Regulatory Commission - Washington, DC 02555 ' l Gentlemen:
SUBJECT:
Draft Regulatory Guide on " Lightning Protection for Nuclear Power Plants," dated August 1979, Task RS 705-4 An Ad Hoc Working Group (3.4.13) ccmposed of members of the Surge Protective Devices Committee (SPD) and the Nuclear Power Engineering Committee (NFEC) of the Institute of Electrical and Electrcnics Engirieers (IEEE), and the American National Standards Institute C62 (ANSI C62) has reviewed the draft Regulatory Guide on " Lightning Protection for Nuclear Power Plants" dated August 1979, Task RS 705-4 (Guide). An extension of time for comments on this Guide from the SPD Committee to November 23, 1979, was granted by Mr. E. C. Wencinger, Chief of Reactor Systems Standards Branch, in his letter dated September 7, 1979, to Mr. W. R. Ossman, Chairman of the Ad Hoc Working Group. The attached comments are submitted with the intent of improving the technical content of the draft Regulatory Guide. In particular, the comments discuss lightning theory to the extent it applies to lightning stroke current magnitudes, lightning protective systems, and ligntning stroke currents bypassing the protective shielding. The discussions and comments are based on the assumption that the nuclear power plant and associated substation are shielded from lightning. For clarification and convenience we have proposed in the attachment the rewording of several sections of the draft Guide. References are included to support the technical aspects of our comments. 90004001
* **%e...db.a $
Power and Life 8001030 - - i
Secretary of the Commission November 13, 1979 Page 2 The technical basis of our discussion also applies to the draft "Value/ Impact Statement." Since specific comments were not made on individual sections of the " Statement," it should not be interpreted as concurrence. Very truly yours,
. , 4
/
m W4.s(' 0; 'Li d (' ul/ %CC '/ ij .. . G. 3. Haralampu, Chairman Surge Protective Devices Committee - IEEE
-..,- . s J. L. Koepfinger, Chairman C6:' Committee - ANSI I T' ~ 6 '
ccc.M% v.c. wee i J. T. Boettger, nairman ( Muclear Power Engineering Committee - IEEE l cc: C. L. Wagner, Chairman, Technical operations Department Power Engineering Society - IEEE 1 I l l l 90004002 l l 1 i l
i I 1 l l i Discussion and Comments on Draf t Regulatory Guide Task RS 705-4, August, 1979 on Lightning Protection for Nuclear. Power Plants by the Ad Hoc Working Group (3.4.13) on Lightning Protection for Nuclear Power Plants Ad Hoc Working Group composed of members of the: Surge Protective Devices Committee - ) Institute of Electrical and Electronics Engineers l Nuclear Power Engineering Committee - Institute of Electrical and Electronics Engineers C62 Committee - American National Standard Inst'itute 1.0 Section B - Page 3 of the Guide 1.1 Line 12 - Replace the words " frequency of lightning surges" with the words " frequency of lightning strokes" This better describes the event and agrees with the definitions in the IEEE Dictionary. 1.2 Line 14 - Replace the words " design basis discharge surge" with the words " design basis stroke" This better describes the design basis. 1.3 Line 17 - Replace the words " lightning-induced surges" with the words " lightning caused surges" The surges being discussed are not induced 1.4 Line 18 - Replace the words " Lightning surges" with the words
" Lightning strokes" Same comment as above 1.5 Line 19 - Complete the sentence with the words "at a lower frequency of occurrence" This clarifies the intent.
90004003 i f
l
'. e
)
. l 1.6 Discussion for paragraph beginning with line 29 Mention is made in this paragraph of the sensitivity of solid I state logic systems to transient voltages generated externally j to a plant. This thought is commingled with the application ;
of surge protection to the power transformers supplying the station auxiliary power. By this action ene would be led to believe that the application of surge protection to the power transformer would prevent failures of sensitive equipment. Surge protection applied to protect transformers or switchgear insulation does not recessarily reduce surges sufficiently to protect sensitive sclid state equipment. This equipment may need its own protective system. This position is supported by work which is being carried out by the IEEE Surge Protective Devices Working Group No. 3.3.6, Low Voltage Surge Protective Devices. This Working Group is producing a Standard titled
" Application Guide for Low Voltage Surge Protective Devices (600 Volts or less)," IEEE Standards Project 769.
1.7 Suggested rewording starting with line 34 and continuing on page 4 of the Guide through line 10 , i "For example, power to redundant on-site safety related electric f distribution systems is typically supplied from the off-site transmission system through a minimum of two power transformers to provide redundancy.
"In any of these transformers, there is a very low probability that a primary to secondary insulation failure would occur for
, any reason. If this type of failure does occur, it could impress the primary voltage onto the secondary windings. To avoid a " design basis event common mode failure" affecting the redundant safety systems, surge arresters or other protect _ve devices shall be applied on the transformer secondary winding syste,m.
"If calculations indicate that the arresters located on the high voltage winding of the transforaer adequately protect -
the insulation of the transformer low voltage winding system, surge arresters are not required on this system. Calculations are treated in the IEEE Tutorial Course Text 79EHO 144-6 PWR on Surge Protection in power Systems.
" Additionally, a lightning caused surge entering the high voltage side of the transformer can propagate to the low voltage circuit through capacitive and magnetic coupling of the transformer. Arresters located on the transformer primary winding will provide secondary winding protection.
"Any of the above surges entering from the high voltage side of the transformer are not expected to damage any of the power plant safety related systems. These systems are low voltage systems at least two transformations from the off-site high voltage transformer."
90004004
_3 I 1.8 Discussion i A lightning strike to one line feeding one of the off-site transformers or a direct hit to the transformer will not disable the plant. The NRC Regulatory Guide 1.6 entitled " Independence Between Redundant Standby Power Sources and Between Their Distribution Systems" endorses IEEE Standard 308-74, entitled " Criteria for Class IE Electric Systems for Nuclear Power Generation Station." These documents state that a minimum of two off-site independent systems have to be available to feed any class IE equipment and that in addition, each system shall have an on-site power source such as diesel power. Therefore, the failure of any off-site transformer should not affect the safe shutdown of a nuclear power plant. *
.i In addition, a lightning surge entering the transformer primary is not expected to reach the low voltage circuits since the impedance between the transformer primary and class IE equipment is high and the surge will have dissipated before reaching the low voltage systems (600 Volt and below).
A direct hit on the lightning protective systems on the plant site will cause the ground potential to rise. It is possible that solid state electronic devices important to safety, will become stressed under this condition. To reduce the chance of . damage, low voltage protective devices can be applied. i t k'
. 2.0 Page 4 of the Guide 2.1 Line 24 - Replace the words " lightning arresters" with the .
words " surge arresters. " j i The former te'rm has been superceded. ' 2.2 Line 31 - Same comment as for 2.1. e 3.0 Page 5 of the Guide 3.1 Line 10 - Same comment as for 2.1. 3.2 Lines 10 Delete the last sentence of the paragraph: "This standard also was no ..." l, We don't believe this sentence is pertinent here. 3.3 Lines 12 Suggested rewording to include additional information which is pertinent:
"The statistical data on lightning stroke and arrester discharge characteristics relied upon for development of the above standards were ecliected in the 1940's. Continuing studies subsequently have verified and added to this data. The most recent studies have been documented by EPRI in the report EL-1140 and by DOE, as outlined in reference 34 of the Guide."
_ . . . 90004005
3.4 Section C-1 3.4.1 Discussion on the Wave Shape It is recognized that in serv _ce, the discharge currents through surge arresters seldom, if ever, have an 8 x 20 microsecond wave shape. Data from field studies indicates that a greater percentage of the arrester discharge currents crest in less than 8 microseconds, and exceed,20 microseconds at half of the crest value on the wave tail.(ll) The use in Standards of the 8 x 20 microsecond wave shape is valid as a laboratory definer for the following reasons:
- a. All laboratories are equipped with surge generators, and wave shaping parameters external to the surge generator, that readily produce the 8 x 20 wave shape for the range of discharge currents considered -
for insulation coordination.
- b. The 8 x 20 microsecond arrester discharge current, in the current range considered for insulation i coordination produces a voltage wave that approxi- l mateg the 1.2 x 50 microsecond wave shape that l 1s used to determine the withstand strength of i insulation. i f
It is recommended that par'agraph C-1 delete any i reference to the wave shape of the stroke current. I It erroneously implies that the discharge current ' through surge arresters is the same as the stroke - current. f I 3.4.2 Discussion on the Design Basis , I i ! The Guide makes no distinction or difference in the terminology of stroke or stroke current, surge or surge current or discharge current. In the electrical power industry and among practitioners of surge protection, these have distinctly different meanings , and relate to different functions and times in the j mechanism of generating and protecting against a lightning stroke or switching surge. In the recommenda-tions presented, this terminology has been clarified and used in its correct meaning.
~
90004006
4 The Draft Regulatory Guide assumes.a finite probability that a 200,000 ampere lightning stroke can occur. It also assumes a finite probability that this 200,000 ampere stroke will contact a protected structure or system and will discharge the full value of this stroke current through a protecting arrester. It is submitted that the second assumption is unrealistic. In fact, lightning protective systems can be and are designed so that the maximum current of a lightning stroke which bypasses the protective shielding and terminates on a protected conductor is limited to a lesser value by design. This lesser value is an order of magnitude less than the possible maximum stroke current. Further, the design of surge and lightning protective systems involves the use of shielding systems, grounding systems, and capacitive coupling systems (as well as discrete protective components such as surge arresters.and capacitors) to reduce, divert, attenuate, and dissipate the stroke and surge energy so that no element in the system is ever subjected to the full value of the incident, stroke current. In. designing the protective system, advantage is taken of the con-figuration, or ability to configure the protected system in order to improve the ability of the protected system to absorb or dissipate the lightning energy and thus render the protected system less vulnerable to damage. i These aspects of the design and application of lightning ; and surge protective systems have been excluded by the l second assumption embodied in.the present Draft Guide. I The IEEE Surge Protective Devices Committee strongly
. urges NRC to recognice the impracti cabili ty of the second assumption, which would have a serious impact on the surge protective industry in developing a new class of arrester for the proposed 200,000 ampere l discharge current, which is over three times the present capability as demonstrated by tests. Such a high discharge current can be prevented by applying the arrester in an overall protective system which has been designed to limit the surge current applied to the arrester to a lower specified value than that of the stroke current.
In making this recommendation, the IEEE Surge Protective Devices Committee fully supports the recommendation in Section 2 of Attachment A to the Draft Regulatory Guide. This recommendation would limit the maximum lightning stroke current which can bypass the station protective shielding by design of the elevations and horicontal separations between shielding conductors and protected conductors. These relationships are controlled to limit the maximum striking distance (stroke attraction distance) l of a protected conductor to about 150 feet. (33) 90004007
l l Control is achieved when, for any possible stroke leader tip location at greater than the specified striking distance (and therefore greater prospective stroke current), there is a protective shield conductor closer to the stroke leader tip than is the protected conductor. l This can be visualized as a spherical ball with radius l equal to the specified striking distance (recommended 150 feet) with stroke leader tip at the center, rolling across the ground and protected structures to occupy every possible position of a descending lightning stroke. Every point touching the sphere wherever it moves will be at specified striking distance and specified prcspective stroke current. Every point outside the sphere will be at greater striking distance and greater prospective stroke current, but will be protected by preferential breakdown and stroke discharge to some point touching , . the sphere. If a smaller striking distance (and smaller prospective stroke current) is selected, this can be visualized as a smaller diameter sphere which can roll down between the shield conductors to touch the protected conductors if the radius is small enough. The basic protective principle is based upon an electro-geometric model of lightning strikes to shielding and protected conductors which has been calibrated against many thousands of mile-years of performance of various transmission line designs. Good correlation has been obtained between actual and calculated predicted results. The theoretical basis starts with understanding the - mechanism of the lightning stroke as it descends from a cloud charge (1) (2) (3) (4) (5) (6). Golde observed that the stroke is attracted at some point in the order of 100 meters from its terminus (7). Golde also observed that the attractive distance was shorter for low stroke . currents, as was also noted by Franklin. He postulated that electric gradient under a stroke leader is a function of the charge,in the leader channel which is in turn proportional to the amplitude of the stroke current. When the stroke leader tip approaches close enough to ground or a structure to develop a critical breakdown gradient of about 5 kV/cm, the stroke terminus is determined (8). Wagner developed a model for the stroke, and predicted stroke voltages, stroke currents and striking distances (9) (10). The model and predic-tions were verified by laboratory long gap breakdown and other experiments (9) (10) (12) (13) (14) (15). Based on Wagner's model of strike distance and prospective stroke current, Brown developed the analytical electro-geometric model of lightning shielding (17). Both models were extensively verified in the late 1960's' under the Pathfinder project which installed some 4600 lightning stroke recorders on 50 transmission lines (16) (18) (19) (20). 90004008
The electrogcometric analytical model has been extended and calibrated against actual performance by Brown to predict discharge currents, line flashovers (shielding failures), and backflashes (21) (22) (23), together with simplification of the analytical techniques (24). The application of this work to EHV transmission line performance has continued successfully (25) (26) (27). Cencurrently, there have been applications of the electrageom'etric model of shielding to high voltage substations, based on the transmission line work and Sargent's analvsis of strokes to tall structures and to open grounc (28) (29) (30). Recently, the work on stations by Mousa (31) has shown the practicality of design to control stroke current to a bus under shield-ing failure conditions by limiting the maximum strike distance. Lee has shown a simplified graphical technique for desigr. application (32).
- There is some variation in the relationship between strike distance and prospective stroke current as-reported by various investigators. However, subsequent to the early work by Golde and Wagner which-was primarily based on theoretical deductions from stroke photographs, there is remarkable agreement among the five relat'lonships in the literature which have been used in predicting transmission line performance. At 20,Q00 ampere prospective stroke current these have a striking distance range from 205.to 281 fect', with a mean of 235 feet and a sigma of 29 feet. At 150 foot striking distance they have a prospective stroke current range of 10,000 to 12,500 amperes with a mean of .
11,000 amperes and a sigma of 1,100 amperes. A very conservative design with adequate margins can be achieved by coordinating insulation impulse withstand with a 20,000 ampere stroke reachi'ng the protected conductor but designing the protective shielding for 150 foot maximum strike distance which would allow a probable maximum stroke to the' protected structure of only 12,500 amperes. It is recommended that the design basis event for lightning and surge protection in nuclear power plants be based on a 200,000 ampere lightning stroke reaching the lightning protective system and a lesser stroke, limited by des!.gn of that protective system (as described in Section 2 of Attachment A to the Draft Regulatory Guida) to 20,000 amperes reaching the protected structure or system. The insulation withstand capability of the protected system and the ' arrester discharge current capability would then be selected to protect properly, and survive without damage, whichever of these conditions produces the most severe surge stress. 90004009
. l By proper coordination and design of the protective and the protected systems, the system designer can control at each point in the system whichever of these design basis strokes will prevail. It is recommended that the changes proposed to accommodate this technical approach, which is the current state of the surge protective art, be adopted in the Guide.
3.4.3 Paragraph C-1, page 5 Based on the above discussions, we reccamend that this paragraph be changed as follows:
"1. DESIGN BASIS LIGHWING STROKE The design basis for lightning and surge prote tion shall be either:
. a. A lightning stroke current of 200,000 amperes
. reaching the lightning protective system; or
- b. A lightning stroke of 20,000 amperes reaching the protected structure by' shielding failure (limited by design of the shielding system) which may subsequently be discharged by a surge arrester."
3.5 Section C-2, pages 5 and 6 - Discussion for sections C-2.1, C-2.2, ; C-2.3, and C-2.4 ' Section 2.0, as now prepared, fails to recognize that the surge protection selected should be designed with two purposes:
- 1. It must provide an adequate protective margin fer the insulation system it is to protect against various types of surges.
- 2. It must perform its protective function and then be capable of returning to normal (r' e sealing) when the surge has been dissipated.
In Section C-2.1 the emphasis is placed upon the use of a surge arrester rated for 100% of normal line to line voltage. This emphasis falls to give recognition to the fact the modern transformer and insulation system for off-site power sources are designed .,t;o have insulation withstand levels which would not be adequately protected by a 1,00% rated arrester. 9000401.0
Section C-2.2 is an apparent attempt to acknowledge that certain systems can and must use less than 100% arresters to properly protect an insulation system. However, it presents an over-simplification of the application rules for surge arrestors. The attempted simplification stated in Section C-2.2 relates to a rule for the acceptable application of a surge arrester rated 80% of the line-to-line voltage. If the statements given in Section C-2.3 are followed, there is no need for the inclusion of Section C-2.1 or C-2.2 in the Regulatory Guide.
! Similarly, Section C-2.4 is unnecessary since ANSI C62.2 and the statement in Section C-2.3 provide adequate guidance for all instances of the application of surge protection for trans-formers and switchgear.
4.0 Page 6 of the Guide 4.1 Sections C2.5, C-2.6, and C-2.7 4.1.1 Discussion - See comments made in Section 1.8. 4.1.2 Suggested' Rewording o'f Section C-2.5
" Surge' arresters with a current discharge capability ,
at least equal to the current to which they would be I subjected by a design basis stroke if exposed to lightning should be installed only on windings of start-up and auxiliary transformers where the insulation ) withstand does not exceed by 'an accepted margin (per J ANSI C62.2) the surge voltage at the winding terminal resulting from a design basis stroke. For redundant systems that do not share transformers, the discharge capability recommended in Section 7 of ANSI C62.1-1975 is acceptable." 4.1.3* Suggdsted Rewording of Section C-2.6
" Surge arresters with a current discharge capability ,
at least equal to the current to which they would be subjected by a design basis stroke if exposed to lightning should be installed at the electrical switch-gear upstream of the feeder breaker connected to start-up and unit auxiliary transformers shared by redundant systems where the switchgear insulation withstand does not exceed by an accepted margin (per ANSI C62.2) the surgo voltage at the switchgear bus resulting from a design basis stroke. For redundant systems that do not share transformers, the discharge capability recommended in Section 7 of ANSI C62.1-1975 is acceptable." , 4.1.4 Section C-2.7 Since the design basis will limit the arrester discharge current to 20,000 amperes, there is no need to consider paralleling surge arresters. Sectior. C-2.7 should therefore be deleted. I
- l. ~
- t 5.0 Page 7 of the Guide 5.1 Section C-2.9 We recommend that this section be deleted. As has been discussed in the revised Section C-1, " Design Basis Edghtning Strokes," a surge arrester when applied to an adequately shielded power system will not be exposed to a discharge current of greater than 20,000 amperes *.
We interpret the reference which is made in C-2.9 to ANSI C62.1-1975 (Section 7.5.1) to a durability design test for a surge arrester. Currently this test specifies an artificially high arrester discharge current of 65,000 amperes. This is used in the design test of the arrester to impress upon its internal parts, voltage stresses which are considerably in excess of those which it would experience in actual use. Thus, there is ; no real requirement for a 200,000 ampere durability test. 5.2 Section C-2.10 5.2.1 Line 3 of the paragraph l Delete the words " design basis" to clarify the sentence. 1 5.2.2 Lin,es 4-7 (Last sentence) Reword the sentence as follows to include the technical basis discussed in 3.4.2 above:
"However, for redundant systems important to safety which .
are electrically connected to these transformers, the surge voltage at each of the transformer terminals and for the discharge voltage of any surge arresters applied on these terminals when subjected to a discharge current of 20,000 amperes, shall be less than the transformer . Insulation withutand by an accepted margin per ANSI C62.2." 5.3 Section C-2.11 This entire secticn should be revised to include wording resulting from the following discussion: Removal of surge arresters af ter a period of field service for retest does not fully accomplish the intended results. Testing at any interval without other monitoring cwans does not preclude the failure of an arrester within the interval between tests. ] The removed arresters cannot be tested as specified in the Guide in accordance with all the performance test requirements of the Design Test section of the applicable arrester standard. ! l 90004012
Because of the limitations of surge generators, and 60 Hert: power sources, the arrester units must be dismantled and re-assembled into smaller prorated sections for a number of the tests. Noteworthy among these are the High-Current and Duty Cycle tests. It is well documented that a prorated section of an arrester can accurately represent, for a particular test, the characteristics of a complete arrester. If the testing of arresters removed from service is required, the testing of the parts cf the larger unit as prorated sections must be an acceptable practice to the Nuclear Regulatory Commission. It must also be recognized that there is the possible consegunece of damaging the arrester during removal and rebuilding the components into the lesser rated prorated sections. The conversion of the larger unit into a number of prorated sections will also L. crease the cost of retesting. It is recommended that an alternative approach be used to l determine the condition of the surge arresters rather than the specified removal and test procedure. *The arrester's condition can be periodically monitored while energized by suitable , devices in series with the arrester. With these devices, reliable information about the arrester's condition can be obtained. Proper judgement of the data from these devices can also anticipate well in advance a potential failure of the arrester.
)
1 There are recording instruments, or devices, with reliable field service reccrds that can provide the necessary data. Surge counters are used to record the number of times the arrester has opers.ted. A Rogowski coil, around the' ground lead of the arrester, that is coupled to a recorder can determine the magnitude and wave shape of the surge current I through the arrester. A current mil 11 ammeter can' determine . l the grading current through the arrester, and if suitably designed, can indicate the presence of conducting contaminant on the surface of the weather housing. A replica of the inter-rupting gap of the arrester can be connected into the circuit so that a judgement can be made of the condition of the arrester gaps. The voltage wave shape of the grading current through a noninductive resistor can be observed on an oscilloscope. The latter gives an indication of corona if present and the magnitude of the leakage current. The periodic recording of the data from su:h methods can'be compared for indications of arrester change or severe duty. A judgement can then be made to remove the arrester. Inspection of the unit removed will serve as an indicator for refining the decision to remove, or leave in place, arresters with similar design and duty records. 90004'013 i
6.0 Page 8 of the Guide 6.1 Section C-3.1 Change the words " Ground wires" to the words " Shield wires." (Accepted terrainology) 6.2 Section C-3.2 We recommend rewording of this section to reficct the discussion in 3.4.2 above:
" Transmission line shielding should ,be designed to limit shielding failure stroke current contacting the conductors to no more than 20,000 amperes."
. 6.3 Section C-3.3 We recommend rewor' ding for clarification and accuracy:
"The footing resistances to ground of the towers for a sufficient distance from the station should be designed low enough to minimize the probability of backflash of the line insulation."
6.4 Section C-4 6.4.1 We recommend title be changed to:
" LIGHTNING SHIELDING FOR PROTECTIOh OF STRUCTURES" 6.4.2 Section C-4.1 We recommend the addition of the following words at .
the beginning of the sentence:
" Lightning shielding, including shield wires or air terminals..."
6.4.3 Section C-4.2 We recommend the beginning of the sentence read as follows:
"The protective shiciding system should be connected..."
6.4.4 Section C-4.3 We recommend rewording for technical accuracy as follows: .
" Lightning protective shiciding systems should bc designed to limit the prospective stroke current of l
shiciding failures (strokes contacting the protected structures) to no more than 20,000 amperes. A striking l i distance of 150 feet is recommended for conservative design." 90004014 _ _ _ 1
7.0 Page 10 - Value/ Impact Statement Comments previously made for pages 3-7 of the guide also apply to this page. 8.0 Page 11 - Value/ Impact Statement 8.1 Line 3 . Change the words "The frequency of induced" to the words "The frequency of occurrence of direct lightning..." for clarification purposes. 8.2 Section b starting with Line 13 Discussion of 3.4.2 made previously applies here also. The parenthetical phrase on line 20 "(if the secondary is not properly grounded)" is not pertinent here. It makes no difference; the phrase should be deleted. 8.3 First sentence starting with Line 22 This sentence should be rewritten as follows: "S ubs tantial effort has been expended to determine a conservative ' design basis stroke.' This work is actively continuing (Refs. 34 and 36-39) . " 9.0 Page 12 - Value/ Impact Statement 9.1 Line 1 . Replace the word " ground" with the word " shield." This is an j accepted terminology. 9.2 Line 4 Replace the word " Conventional" with the words " Protective shielding and lightning rods..." 9.3 Paragraphs e and d starting with Line 8 Replace with the following:
" Installation of surge arresters to adequately protect switch-yards and substation equipment in accordance with good engineering practice and existing standards."
9.4 Line 12 This line should read as follows:
" Installation of low voltage surge protectors on. .
10.0 Page 13 - Value/ Impact Statement Section 1.3.1 - Fourth Paragraph Replace the words " surge characteristics" with the words " lightning stroke characteristics /'and the words " surge amplitudes" with the words " lightning stroke amplitudes." In addition, the end of the first sentence should read "... and frequency of occurrence." The equipment insulation tolerance to voltage surges is defined in several ANSI and IEEE Standards and Guides. The NRC should identify areas needing attention. The IEEE Technical Committees may be willing to provide the technical expertise. 11.0 General Comments on the Value/ Impact Statement - All our discussions and comments on the Guide are also applicable a to the Value/ Impact Statement. Proper correlation should be made. 90004016
.J REFERENCES (1) B. F. J. Schonland and H. co1 P ,is , " Progressive Lightning," Proc., Royal Soc. London, Vol. 143, pp. 654-74, 1934. (2) B. F. J. Schonland, D. J. Malan and H. Collens, " Progressive Lightning II," Proc. Royal Soc. London, Series A, Vol.152, pp. 595 . 625, 1935. (3) D. J. Malan and !!. Collens, " Progressive Lightning III, The Fine Structure of the Return Stroke," Proc. Royal Soc. London, Voi. 162, pp. 175-203, 1937. (4) B. F. J. Schonland, D. B. Hodges and H. Collens, " Progressive Lightning V, A Comparison ,of Photographic and Electrical Studie,s of the Discharge Process;" Proc., Royal Soc. London, Vol . 166, p. 56, 1938. (5) I. S. Stekolnikov, "The Nature of the Long Spark," I:datel'sivo Akademii Nauk USSR, pp. 1-272, 1960. English trans. available from Foreign Tech. Div. , Air Force Systems Command, Wright-Patterson AFB, Ohio. ' (6) C. F. Wagner and A. R. Hileman, "The Lightning Stroke II," AIEE Trans., Pt. III (Power Apparatus and Systems) , Vol. 80, pp. 622-42, Oct. 1961. - I (7) R. H. Golde, "The Attractive Ef fect of a Lightning Conductor," J.I.E.E. , vol. 9, p. 212, 1963. , (8) R. H. Golde, "The Lightning Conductor," J. Franklin Inst., Vol. 283, pp. 451'-477, June 1967. (9) C. F. Wagner, " Relations Between Stroke Current and Velocity of the Return Stroke," IEEE Trans. on Power Apparatus and ' stems, pp. 609-17, 1963. ( (10) C. F. Wagner, "The Lightning Stroke as Related to Transmission Line Performance," Parts I and II, Elec. Eng., May and June, 1963. (11) K. Berger, R. B. Anderson, H. Kr6ninger, " Parameters of Lightning Flashes," CIGRE Committee 33, Electra No. 41 (12) C. F. Wagner, A. R. Hileman, " Mechanism of Breakdown of Laboratory Gaps," AIEE Transactions, Pt. III (Power Apparatus and Systems) , vol. 80, pp. 604-22, October 1961. (13) C. F. Wagner, A. R. Hileman, " Surge Impedance and Its Application to the Lightning Stroke," Ibid., pp. 1011-22, 1961 (i eb.1962 section) . (14) C. F. Wagner, C. M. Lane , C. M. Lear, " Arc Drop During Transition From Spark Discharge to Arc," AIEE Transactions, pt. III (Power Apparatus and Systems), Vol. 77, pp. 242-47, June 1958. (15) C. F. Wagner, " Lightning and Transmission Lines ," J. Franklin Inst. , Vol. 283, No. 6, June 1967 90004017
S 4 REFERENCES (16) H. R. Armstrong and E. R. Whitehead, "A lightning stroke pathfinder," IEEE Trans. Power Apparatus and Systems, Vol. 83, pp. 1223-1227, December 1964. (17) G. W. Brown, "The Electrogeometry of Shielding Against Lightning," Ph.D. dissertation, Illinois Institute of Technology, Chicago, Ill., 1967. (18) H. R. Armstrong and E. R. Whitehead, " Field and Analytical Studies of' T- .nsmission Line Shielding," IEEE Trans. Power Apparatus and Systems, V;1. PAS-87, pp. 270-281, January 1968. (19) Gordon W. Brown, Member, IEEE, and Edwin R. Whitehead, Fellow, IEEE,
" Field and Analytical Studies of Transmission Line Shielding: Part II,"
IEEE Transactions on Power Apparatus and Systems, Vol. PAS-88, No. 5, pp. 617-626, May 1969. (20) E. R. Whitehead, " Final Report on Edison Electric Institute Research j Project No. RP-50, Mechanism of Lightning Flashover on Transmission Lines," No. 72-900, EEI, 90 Park Ave., New York, NY 10017, (1972). (21) Gordon W. Brown and Steven Thunander, " Frequency of Distribution Arrester Discharge Currents Due to Direct Storkes," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-95, No. 5, pp. 1571-1578, September / October 1976. (22) Gordon W. Brown, " Lightning Performance II Updating Backflash Calcu-lations," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-97, No. 1, pp. 39-52, January / February 1978. (23) Gordon W. Brown, " Joint Frequency Distributions of Stroke Current Rates of Rise and Crest Magnitude to Transmission Lines," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-97, No. 1, pp. 53-58, January / February 1978. (24) Gordon W. Brown, " Lightning Performance - I Shielding Failures Simplified," IEEE Transactions on Power Apparatus and Systems, Vol. ' PAS-97, No. 1, pp. 33-38, January /Feburary 1978. (25) D. W. Gilman and E. R. Whitehead, "The Mechanism of Lightning Flash-over on High-Voltage and Extra-High-Voltage Transmission Lines," Electra No. 27, pp. 65-96, March 1973. (26) E. R. Whitehead, "CIGRE Survey of the Lightning Performance of Extra-High-Voltage Transmission Lines," Electra, 33, pp. 63-89, (1974). (27) E. R. Whitehead, " Analytical Speculations on Improved Electrogeometric Models of the Lightning Flash and Transmission Line Environment, with Addendum and Appendix. IND 12 and 12a SC33, WG 33.01, (1976) 90004018
w - m+.+4E .i.. h 4-.- - - .a.n+ w.A.meu e =wa .
- q=+ - e- 4 s , .i.-m . -,-=- - a - m - -r .. e.w.
REFERENCES . (28) M. A. Sargent, "The Frequency Distribution of Current Magnitudes of Lightning Strokes to Tall Structures," IEEE Trans. PAS-91, No. 5, pp. 2224-2229, September / October 1972. (29) M. A. Sargent, " Monte Carlo Simulation of the Lightning Performance of Overhead Shielding Networks of High-Voltage Stations," IEEE Trans., { Vol. PAS-91, No. 4, pp. 1651-1656, July / August 1972. ; (30) H. Linck, Shielding of Modern Substations Against Direct Lightning Strokes," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-94, No. 5, pp. 1674-1679, September / October 1975. (31) Abdul M. Mousa, " Shielding of High-Voltage and Ixtra-High-Voltage ~ ' Substations," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-95, No. 41, pp. 1303-1310, July / August 1976. (32) Ralph H. Lee , " Protection Zone for Buildings Against Lightning Strokes Using Transmission Line Protection Practice," IEEE Transactions on Industry Applications, Vol lA-14, No. 6, pp. 465-470, November / December 1978. . . (33) Ralph H. Lee , " Lightning Protection of Buildings," IEEE Transactions on Industry Applications, Vol. lA-15, No. 3, pp. 236-240, May/ June 1979. i 90004019 l e . I
1 rod to. plane gap upon applicatioi of a 60 Regarding the development of the spice the earl > stages of the channel formation cycle voltage or an irnpuhe having a slow charge, Park and Cones stated that 'An and its presence is not apparent. front and long tail The higher value of 160 analysis of a large number of records ob. In reply to the coininent made in the lut l kv per f int or 5.000 voit, per ein represents tained with slow!y ruing sures indie sted paragraph of Mr. IIagenguth's discus, on.
'he d tr acteristic 4 a r.<l.r< =1 gys to w hich that the peak curren: a.as approxi n.ttely ]
.t ne;;it n e p..te nt t.d is .1pp.' icd i t w ould be it does not appe.1r that a glow divhars;e or i proporuunal to the act J d value of voitue pilot. leader without some sort of coaducting interrit;::g :f dat t pertavang to a rod plane at the instant the disc harge started ' core ichanntil would posses.s audiesent con-g ip were obtained in a range 'I 400 inches Therefore,in an ambient of Lw free efectron for whius the applied voltage i3 .t ne;citive ductivity in the form of a cylinder 1090 concentratisin and woh the 2; piication vf a or ;'U.000 fert in length an.1 1U0 feet in di-skw. fron t wave The se data cou!d be stee;> voltage wwe, the crut value of the cornp ired directl$ with the low grxlient of ameter, to supply the current required to voltage w ave -s . tttamed brf are tnggering provide the progrewin g corona lischarge I,5 0 n,its per crn obtained wit h i p.muve occu r s hat if the cweentration . if free ( space ch.ir; in front of the !cader Fur-aase Fig.14 of refereta e 1 of Mr. H .cn eh c t ron s a high. tria ng may . iccu r . .n ther r ore d the Nader consir 'I of only such guthidncussion prm: des iorrm informauon the nsmg portt >n .f t he w as e with a cor- a glow di,c n a r;e. the gradient per unit
. ocerning rrw1 rod gtps w hich indicates responding re;fueren in crest value of the lengt h must be approximately W) voits hueanty up to 180 inches and a gradient curre n t . It is to be presumed that a corre per cm. The . trop alone m such leader of I
rd about 5x00 volts per em. spondin g lenginemng of the current wave Ph.uoO feet length would be 5 X lui vo;ts l The question raised by Mr. Ifagenguth would ensue. According to our theory of This wouH require a de;%ition .f charge betr.n s that we were net su f.ae n t!: e ear brendow n, the substar.ti.ii lesel >r:ent < f along the st roke enannel that inc rea ses l m the yncral expuiti n of the paper We the space chare is t precedent to t he de. linearly with he <ht rie re,u; tant current i t rie d i cons ey tnat sprk wer ui the gap vel._ pment of dt anr.r! The curre*it required
.n ur, m two phases. Srn tb ies elopme nt at the earta. as t:ie return <troke t apped to devebp the sp ic- Eiira is small in the-e ch.n ges progt e s ,is e ly, aou!11 resa;t vf t v 3p4ce c c tr e < cor. na h s irge),and compartwa with the ,iNrt -circuit eurn e 1 secor o b the de velopme n t of the channel in a current itene arth t hat w ould increase !
of the sure pnerator 4 hen uiumate bred. progresdneo oth time up to about 100 j i high conducting are plasma L Only the down occurs Therefore, when the current usec and in magmtude would be many times
- irst ph.i3e develops below critical voltage. shunt is adjusted to read the short.ciremt i
.\bove critical voltage both occur in se- the recorded values Thus we are of the i current, the space charge current is swamped opinmn that a conducung core 'nust exist (
1"" ' " he channel formauon currents even in within t he
- Wr fe @ /]
_ _.. . u -
~l _ _ . _ _ _ _ . _ . _ . _ _.
I
- l C l ll ther progress of the streamers, condi-
&6 Tlne Liontu*lno a g OtTOKO II tions just in advante or the tip of the channel beame conducive to the initia- ,
tion of a channel or are plasma at this C. F. WAGNER A. R. HILEMAN l FEuCT AEF YMER NEE point. This new channelin reality merely constitutes a further extension of the i leader channel. Each new channel spurt st rts with a relatively low velocity that l INtitleA PREVIOB PAPERJ smnbrly which wdl be caHed the eurona sheath 1 the authors underNR to syn. The diameter of the channelis ontv about ' I"U"ws a curve with time that is strongly thesire certairi characteristics of the li.;ht- 2 mm (mdhmeters. and its drop ahout 50 conca ve upward This continues until nmg <troke by applym; and extrapolatmg or 60 volts per em ' centimeter 0 It has the channel catches up with the boundarv the res ults of !aboratvf experiments, characteristics of an are piasma with verv of the corona -heath The channel can. They were supported in this ehrt by hi.;h temperatures and may be highir m t progreu into ntgin air in the form of data concerning the 'ransient character luminous The diarneter of the corona a highly conducting plasma and. there istics of arcs2 and the properties of corona envelope may be abi<ut 100 feet and may fore, eases In the meantime the corona within cylindrical :he;!s A companion extend about 150 icet in front of the chan- Sl"***" '"ntinue to pro.;tes tram the raw n ths issue. dtseum the proper nel The internal gradient of the corona new tip of the channel and the whole ties of laboratory-praduced + arks and sheath lies between 1000 and 1050 proes is nprated The phogaphic the present paper applies this mimrmation, volts per em It has characteristics of a studb of Schonland' and his asociates together with additional dat; concerning glow or cornna dischar;e, its temperature reveal this rapid extenavn of the channel natural lightning. to a more detaded con- is low . tt is pierced by streamers; and as a short step d verf hid bnHiana sideration of the lis;htning stroke A new con uderable di$culty is sometimes ex- And with respect to the development of mechanism of the lea ler steps is pre perienced in ph~t-graplung it the channel (which thes term streamersi sented. Also, a thew of the cerv im As the channl ..i 'he :cader of the 6nt they sa y, " Derinite evidence that the portant events that occur durmg the component of a stroke reaches a partic- sinamers ManneN trad dowmrd is. early stages of the return str< ke W e!ud ubt pomt er i, momentarde arrested and hownen atTorded by the broadening of dat ed. streamers for;e ahead into virgin air the upper part oi thetr tracks. These streamen +rm the corona sheath &corWng to ahonland, the lengths of General Description of the Stroke and as thev proceed distnbute a space the steps vary between 10 anji 80 meters charge that he characteristtes simdar to a with a modal value or about ao meters or
. Before di3cussmg the various phaac, corona h3 charge and to the 3 pace char;e -
or the stroke a general description of $per 01448. recornrnendert sy the AIE E trant assuetated with the f"rmanve stage of the mouon ena omnnmon comnuu~ ana worono the stroke without detailed substantia- breakdown vi long gaps As the space [",7,,%n t 4t gl nli of m tion wdl be presented The hypothem charge devele the potential liderence cemai omna menne. Ne- o % ns u. pictures the !eader as compmed of two across the corona sheath has an mereasing [$[ [ ' NZS$,' ["r((j',77 parts 2 very th;n good conductmg core, etTect in estrammg the prNres ai the dis- W which will be caned the chanm l. preceded enarc- But. hefore the charga can he C F Lm e and t R n rt.a m ar e % .< h wh Weso n cou e Elecu te Cor por u t mo p. t t +. and surrounded by a negat n e tace c har:e 1.a s t come ud!:. e:In tive m heedng the :ur-
%, o e ,
C % g na r (!ilm a c 1h, l: g w.q M .. (l n w w.g l%; 90004020
d ~
,1 ',
150 feet. The corona sheath advances across the lateral sections of the corona f with an average velocity of between 1.5 X e sheath. Since the charge at the very tip # 10' and 8,0X10' em per see (second) or constitutes such a small proportion of the ' " between 0.0005c and 0.0025c, where c total charge, it cannot have a very great is the velocity of light. The time re- ! [E' , influence upen the potential of the con- Cl #[ ,,, quired to develop the corona sheath be-f ore it is again overtaken by the channel, ducting channel lying on the axis of the downward leader. For the moment con-gj n i
~
o i usually ranges from 30 to 90 usec (micro- sideration will be limited to the essentially j seconds), and the velocity of the step cylindrical portion of the leader. It is o % g exceeds 5X10' cm per see or 0.16c. shown in the companion paper that in g
' },g
- 9 However, more recent measurements of laboratory discharges, just prior to break-
, ~
{ electric fields next to the earth,8-7 in- down, one function of the space charge
?,, -. ] p,a .. ,
i dicate that as the earth is approached the .'" ,.Jr' appeared to be to equalize the electric l time intervals between steps become gradient so that the uniform gradient
'soo -
l [Q j ' smaller and attain a value of 13 usec.
]J equalled the average gradient. '
j Schnnland and his associatesta-in re- _
' In a cylindrically symmetrical space [
ported uncompleted leaders which ceased to :velop before reaching the earth. charge, that distribution which leads to a uniform radial field is one for which the b [~. They also found that the intervals be- A2., density varies inversely as the radius. z
.ym tween steps aim the length of the steps Thus z r -_.M;"
sometimes remai ted constant during a considerable portion of the leader path. 9.I* in coulombs per em' While the foregoing description applies r (1) U 3 co*
*OErA:==# "N t to the more common cloud initiated
- stroke, a smular phenomenon occurs with where r is the radial distance from the axis i l
earth.imtiated strokes. Hagenguth and Andersonu presented a photograph of a and .I, I is a constant that represents the d% u 1 m u is & w M Nd7 ih&N l 4
.e .. . . .
stroke that was imtiated from the 1,2.,..0- that if the charge extends to radius *, with 3 i t foot Empire State Building, it ex-E, the radial field m. volts per em, q, the < hibited very pronounced steps that oc-
- 0*#
" E*# **
- curred at intervals of approximately o.a. of the channel, and I',, the total radial r-C i
g usec. The explanation for the formation , g ' e of steps must, therefore, be independent
- 18X 10u of polarity except in degree. r* g, m em e
g' (0 , The step process of the lightning stroke, 'o
- g. is, in some respects, simpler than labora- and
, j
( tory produced discharges. The forma- , tion of the space charge in the case of the V,,- 18x10" g, in volts (3) [ stepped stroke always emanates from an 4 are plasma constituting a copious supply Such may be the distribution of charge I of free electrons. There is, therefore, no ar und the wire of a wire plane electrode ! - 3"8* pri r t breakdown. _i [ statistical time lag during which the
! initiation of the discharge awaits the b*.** the leader tip is always in a state
* . ,ma%cw;,4% w es 1
propitious positioning of a free electron. Iincipient breakdown it is not unreason-i, The process .is not comp..ucated by the able to assume that after the tip of the necessity of considering the development channel has passed a particular point the Fig.1, Approximate form of corona sheath of st.eamers from an opposing electrode distribution still remains somewhat the at it, most extended point beyond the channel same. Furthermore, at breakdown the J. until the leader nears the earth, as the head of the channel has progressed g gradient for a negative electrode is about to 1,000 feet above the earth [ Analysis of the Stepped Leader 9 (m/g g o g 3-4, POTENTIAt. AND CHARGE DtSTRIBUTION - d A -- I -. 3 ' l As a preliminary stepin the discussion Table 1. Determination of Leader Potential, harge Densities, and Corena Sheath Enve! ope g . of the nature of the step mechanism, it r
'l will be necessary to establish the stroke y potential and the general natare of the Contribution to Potential at These Pointe x tos Volta A one to chare. Densierin Thia charge distributed along the leader. Row section 2 b Since the potential of the channel is as-1 3 4 5 j stuned to be the same along its entire 3 length, then the sum of the drop across
{ {
.a.
4s 3.2. s.o..
, ip ap .y h . .1s.o., 1.3.. 0 14 the corona sheath and the drop from the {,
Corona sheath to ground must be the /. .
. F, is 108 vMts. . . . . .
i . Q 23 8. Q
't5 8.
;*0 [
.42 4
;3N, 7$ $
.42.0 . .44.32 same at any point. To determine this l( es s potential some assumption must be made L j",,*,0 ,
N"'","f"'" " Ya , a s jo
.= in x ios vat .
jk with regard to the distribution of charge f. . . V in x los watts . .
. 32.0.
55.8,
. 21.3.
47.1. 12 2...
.54 6 ml
, 52.1. 43.2 to
*, OCTOBEn 1961 agner, Hileman-The I.ightning Stroke--II 623
}Q
I 0,000 volts per em, and since no branches charge on the other elemental cylinders Ahead of the channel tip the space of the stroke are init!ated from behind the can be obtained. This also applies to the charge density and the electric field are tip, a somewhat smaller gradient of 7,000 negative image charges. Rows a to e thus probably more intense than in a radial i volts per em may be assumed to exist show the contributions to the potentials direction behind the front. The electric there. Therefore, if in any section q is at the mid points of the tive sections due field ahead of the channel approaches
, assumed then from equations 2 and 3, r, to the chstges in themselves and theother and V.,, are determined.
the critical value of about 0,000 volts per four sections. Row f is the sum of the em at which the enannel extension again 5 An instant in the prugre>5 of the leader individual contributions. Adding the commences. will be chosen, shown in Fig.1, at which internal potential drops given in row i to the channelis just about to begin another these values gives the potentials in row j Ttste row 9.scr Cmos FonstAttos ut>
; spurt in its advance. The enrona sheath f which represent the potentiah of the Vet.ocny or ITs FORotXrloM alom; the channel and in advance of it is points on the channel just opposite the
, fully developed for that oarticular step. mid points. These should be equal in little information is available from Only that part of the vertical path below order to satisfy the condition that all laboratory data upon which to base these
, an altitude of 10,000 feet is considered. points on the channel have the same quantities. In reference 3, it was shown
% hile there may be other codetting paths Potential. Thus, for this conditmn, the that from the Park and Canes data" on sphere to-plate< 'aps, with the sphere with.m the cloud it is usumed that they potential of the channel is about 55X 108 have no.effect upon the phenomenon oc- ..
volts. In addition to the chaqes in the negative, fcr an 11.5-cm gap the time for .i curring at the tip. It has been shown by cylinders, an approximately hemispheneal t e corona current to reach zero was 0 3 Schonland, Hodges, and Collens d that bowl of chat;e at the end of the leader, sec. This corresponds to an etTective { the total charge on the first leader when it velocity of charge formetion of 0.0006 c.
, must be included in the computat%ns.
l is fully extended is about 60To of the The total char;e within a sphere whose With the sphere positive the velocity was
; char;;e m the section of the cloud on,gmally 0.0013 c. The Hagenguth, Rohlfs, and l charge density varies as A/r is just 1/2 of , tapped by it. Calculations mdicate that Degnand data for a 200-inch rod. rod gap I the total charge within a cylinder whose j such a proportionate part of the charge in length is equal to its diameter and whose with the anode grounded gave a time of j the cloud, because of its distance, has a formation of 9 usec, which corresponds density is A/r. On this basis it can be to a velocity of 0.0009 c.
negligible etiect upon the phenomenon estimated that, in this case, the contribu-occurring at the tip. The tip of the Schonland's modal value of 150-foot tion of this charge to the potential at steps indicates a time of formation of the channel is assumed at this instant to be 1,000 feet above the earth. The leader point I would just about be equal to the neptive space charge of 50 usec. Since contribution of the char;e in element 1, as assumed to be div @ d into tive the average need travel only half the dis-cylindncal sections with a hemisphencal which is 4.3X 104 volts. This would re-tance, this corresponds to an average quira a slight modification in the charge velocity of formation of dome at the base. The four bottom densities and the radii of Table I. With cylindrical sections an I the hemispherical these computations as a background. the IM'do 49 dome are shown in in 1. 'Ibe bottom 43 in' em per sec '5) curved shape shown in Fig. I was drawn 2 X M x 10 -* cylindrical section is 50 feet long and the as being representative of the form of the others in succession are 200, 2M.1,500, = 0 0015c chaige volume surrounding the channel and 7,000 feet respectively. An intinitely n which the internal field is 7,000 volts For the positively projected steps" from small diameter, perfectly conducting per cm. This is only an approximate re the Empire State Building, cited earlier, a core or channel is assumed to extend from stdt but since the theory, proposed here, 50-foot step, with a time of formation of point A up through the column-is not critically dependent upon the shape 25 usec, gives a velocity of formation of In Table I, the results of a cut and4ry and distribution of the space charge it g.gg computation to determme the char;e dis-was deemed surTiciently accurate for the # x10' 3 cm per see (6) tribution and the potential of the core ,
~
purpose at hand. The average density are shown. The charge densities in the of charge over the bottom 500 feet is about " 0 # 1' five sections are shown in row g. Row h gives the correspondmg rado of the O X 10 ^* coulombs per em. If this The lightning and laboratory data com-charge is drained to earth as a wave of sheaths and row : the corona sheath po- pare favo ably. Not all of the step inter-tential drops. current at a constant velocity of '30"o val time can be attributed to the forma-that of light, then in accordance with The inset of Fig. I shows one of the tion of the space charge as a portion is equation 1 of reference 1, the dischar4" also required for the formation of the elemental cylinders into which the corona would develop a current of 40.000 ampere 3 channel sheath is divided. To determine the or at a velocity of 10"o, 27.000 amperes. potential of a noint p on the surface of the Schordandu arrived at a value of charge VELoc!TY OF CHMGF.t-sheath midway along the cylinder, the density of SX 10-4 coulombs per em from volume charge can be assumed to be con- So far the comparison of the character-entirely different considerations, such as
. centrated along the axis of the cylinder. istics of the step of naturallightning with the modal values of charge lowered in a The potential of this ;)oint due to }he those of Wratory. produced sparks have complete stroke, the ntunbcr of compo. showtt y fying agreement. However, charge on the cylinder ts then, from Iig.
34 of reference 12, nents in a stroke and the length of the when one compares the velocity with stroke. More will be said of the develop- which the channel advances, the agree-gg ment of the actual current to ground. ment ceases. It was mentioned in refer-4 = 9 X 10"D in volts (4 )
- r. It is of interest to observe that for this ence 3 that in the Park and Cones ex-case the potential of the channel falls periments" the head of the positive chan-where q, is the charge in coelembs per within the range of 10' to 108 volts ac. nel starts with an initial velocitrof 4 X 10' em. By similar expressions given in refer- cepted by the majority of the workers in cm per i7 3r 0.00013 e and increases ence 12, the potential of point p due to the this tield. gr.W j .3 'ttst and then more rapidly, 62.s \\*agner, Hileman-Tiie Lightning Stroke-U Octonen 19G1 1
90004022
, , ?
atte.ining a terminal velocity of about fortns a pencil of zero electric field along a velocity of 0.15 c. While this may not be 1 08 cm per see or 0.003 c. The negative line where previously the gradient had the actual travel time, it does correspond channel seems to travel with a greater but been constant at about 9,000 volts per to the time indicated by Schonland. In i unknown vebcity. Since similar thermal em. If the extension of the channelinto any case, it can be used later as a basis procenes must be involved the velocity of the space char;e is very rapid and if it is for reference. At the very tip the current both polarities should be of the same order assumed that the surrounding space reaches a maximum value of 7,000 am- ; of magnitudt Schnnland" states that charge cannot change rapidly, to achieve peres. At any point the product of cur. the steps of the downward leader com- the condition of zero gradient parallel to rent and duration of the dow is constant. plete their passage in 1 usec, which for the axis of the extended channel. charges Now consider the case in which the the 150 foot step to which he was refer- must be induced in this part of the chan- extmion is connected to the channel at A ring, involves a eelocity of nel that produce 2 grad!. nt parde! to the as shown in Fig. 4. While the extension
,g g channel that is just equal and opposite to is traveling downward, the disturbin; in- 4
- 4. > X 10' em per see d 15 c that of the 6 eld before being disturbed by Scence of the sudden propagation of the g the projection of the channel. To form an idea of the char;es and currents in dernward step is also felt in the conduct.
in leader behind point A. The velocity [ i This is a serious discrepancy but as the the extension, two conditions should be with which a disturbance travels al;n.; the ' analysis procteds it will be explained assumed. First, that no current is fed are p asma appears to be degndent upon . satisfactordy. into the extemi. n fri m .l. that is, that the the previ .us intensity of ionization of the extension frem .! consists of an elungating plasma, as is evidenced by the velocity of j DEVELOPMENT oF CHWEt, C1.WENTS conductor insulated from the main chan- propagation of dart leaders and return j netat A. Second, that this connection is strokes. In this case assume that the L Fig. I shows the column of the light-el sed and that the channel and its ex- disturbance travels up the channel at a h ning discharge at an instant when the
#" "#I 7#" # "5' space chargs in advance of the arrested vc!"eity of 50'"o that of light. Thus, dur-channel has reached its point of greatest For the 6rst case, the problem is the ing 1ssee whde the channel is moving advance. Conditions just in advance of deter'r.ination of the charge distribution downward 150 foot fram A to C, the in- [
the tip, point .1, of the channel are at this al ng a cylindrical conductor whose length duenced portion of the channel is the f' instant propitious for a further advance is large with respect to its diameter, whose 500 foot above A. Let it be assumed that of the channel. It seems as though thesa total charge is zero, and whose electric the radii of the leaders on both sides of ' conditions are related to a certain com- gradient along the surface of the conduc-5 bination of high chari;e densty and 6 eld y p ram M @ ads is knma Assum. It certainly involves the same relations ing a step whose len.;th is 150 feet, whose B 1 as the transition from a glow discharge graWent is 1000 volts per em, and whose f . to a high conducting are plasma, for this r dius is 0.0L foot, or 3 mm, an approxi- C 1 mately triangular char;c distribution re- - is the essence of the phenomenon involved. suits. as shown in Fig 3(A), whose den-f ] ; 1 j This is essentially thermal in nature and / 3 j involves considerations of high energy sity at the leading tip is -15X 10" cou- /
@ j l 1 mbs per e and at the trailing tip of 3 j, concentration. The conditions condue:ve 1.5 X 10 4 coulombs per cm. The re- / 4 p l
to this transition are not well understood. suiting density is not sensitive to varia-f ; 4 \ 4 In the companion paper it is indicated ; {, . that when the anode is far removed from l'"ns tn radius; thus, decreaung the ra. #; ; . the cathode, the critical gradient at which dius to 0 00'i foot increases ti e density pi 7 a space char;e develops into an are about 10% and increasing the radius to M ' i y[ plasma at the cathode is in the order of Om foot decreases the density about g ' O,000 volts per em. Fig. 2 shows another $t AS the head assumes intermediate j{ p sitions m its travel, the distnbutions view of the tip oi the leader. The density f , g; j e ntinue to be approximately triangular g}r7 g.- g/( of the dots is intended to be suggestive l ;g
** I '
of the charge density. From this condi- "Y "il " t-f- O;j;{fj twn it may be assumed that the conduct. In justi5 cation of the diameters of the channel extensions used here, reference
; @ggD L J ing channel extends itself from position A is made again to the work of Hizham and
$ /
j to position C and as it extends into the k } space charge it quickly attains a low volt. Meek.t* which shows that the diana / \ [y[ 25! f j[ a;e drop. Justi5 cation for this statement a 500-ampere are grows from 1 mm at 1/4 \ _/b, / i. is provided by IHgham and Meek, who usec, to 2 mm at I usec, and 4 mm at 6 /
\ / '" '
(( demonstrated that for currents in the . range of 60 to 500 amperes that rose to usec. This curve is reproduced in Fig. 4 of reference 2. h a d - N jf p cre.st in 1/4 asec, the drop reduces to 150 Current must Sow within the channel I volts per em in 1/2 us ec. This drop is extension to achieve such charge redis- +: b smallin comparison with the e!ectric deld tribution. If the tip of the extension Fig. 2. Tip of the downward leader of the *
- of the space charge, which had been as- moves with constant velocity then, as it first c mp nent ia ser ke, showing the , {
'# * " " ' ' 8' * "
sumed to be 0.000 volts per em. The high reaches a certain point the current rises the corona sheath and the charges that d tmttal drop may still be s.igmncant m its suddenly at that point and remains con' must be induced in the channel entension r efect upon the phenomenon. However, stant until the channel has reached its and in the channel above the channel tio, ! for the moment assume that the drop maximum extension. These relations to, ieye,ai no,;tions of the channel extension, along the channel extension is zero. are shown in Fig. .E B) for cn assumed assuming that the channel entension occuri [ The projected channel as it advances travel time of the head of 1,see with a rapioly k i Wagner, Hileman-The Lightning Stroke-fl Ocrc>nr.n 1001
} G25
4 A B Fig. 3. Chars for additional charge to annut the electric I and currsnt distri. field existent before the development of bution in the chan* the leader. cua aeNr net extension assum. cNaast o c N serr ' m Au* tats ing that no charge .The enterion that determined the ma;- IN CoV L / C M " . " 4000 8000 is drawn from the 8 '
' * '- o electric lield parallel to the channel ex.
P channel behind its [ a
'\'
tip during the tension produced by there charges is interval that th, equal and opposite to the 6 eld into which 4 n 3" ' M extension is taking it is projected. This results in a zero
\ place electric 6 eld parallel to the axis of the
,, ' * c 8 3 \
'S
- lo * * " e.stension. No such cancellation exists I na !, * ,
for the radial neld along the extension of
$ ', the channel or for points in front of the
[ T
]
g
, extension. If, for example, the charge .
% ; density attains a value of 1.5 X 10-4
\
go : 3 coulombs per em that is indicated in Fig. io-3, the electrie fiehl 4 mm to the side or 2 mm in fre.nt of the tip would reach the l following fantastic value; see equation 5 of Fig. 5 of reference 1: A are 0.01 foot, or 3 mm. For the to an are and also by the transient char. moment assume that this movement is so acteristtes of an are as the current tends to Er = 9 10" ,q= 9X 10n X I.5X 10-* rapid that the space charge contributing '8) increase rapidly. In Fig.11 of reference a 04
- to the deld cannot change in intensity in 2,it is shown, that for an impulse current the immediate space surrounding these ,,3 yn n;' volts per cm wave that rises linearly from zero to 10.-
lengths. In this case the char;es induced 000 amperes in 1/4 asec the voltage drop on these lengths are approximated by the This merely means that in the attempt at 1/4 usec is 3,200 volts per inch or 1,250 solution of the electrostatic problem, for to maintain a low axial field, charges of volta per cm. For shorter instants, with which the potential from A downward in- great value develop which in turn produce the same rate of rise, the drop is even profuse ionization in the space surround. creases linearly to 150X30 4SX9.000 or higher This is merely indicative of the 41 X 104 volts at C and upward from A in.; the channel. This is merely another voltage drops that may occur with very t'orm of corona or glow discharge. The the potential is the same as that at A. rapid changes in current. Because of the The average potential of the %0-foot radius of the corona sheath would expand rapid velocity of propagation above point to about 10 times 2 mm, or 2 meters. length must shift subject to the condition A, the reference ;>oint in Fig. 4, and the that the total charge is zero. Fig 4(A) But this would not vitiate the argument longer distances involved above J, it is concerning the axial gradient. As the shows such a distribution of potential. Likely that the are drop will also The determination of charge density, in corona expands into the original space intluence this portion of the circuit and charge, equalization of charge density terms of known potentials, is rather diffi- otter additional impedance to the flow of cult without recourse to a digital com. results which in turn decreases the 6 eld current downward from A. Actually, originaUy responsible for the ejection of puter, but the inverse of determining the are drop tends to suppress the char;es the channel In this manner the charge potentials, in terms of charge densities, is in both circuits to some extent, because density below A in Fig 2 is gradually relatively simple. In Fig. 4(B) the dotted whatever drop does occur reduces the need transformed into a charge volume density lines show an assumed charge distribu. tion of t.he nature expected in this case. The solid lines show the corresponding A 500 3 - C .- D potentials. It is found that the potential , distribution approximates in shape the {
.aoo ll ,
one under consideration. From this - 1: ~ solution the desired charge distribution produced by the known potentialis given 3on} - Il ! ! I o a SEC in Fig. 4(C). ,
,m0 ; -
This curve of charge density is replotted 's in Fig. 2 by the curve abc. Other curves
,co @
t
\ os for one fourth, one half, and three fourths a
! , o 25 of the travel distances are also shown in ,__o i
* ' Fig. 2 and in Fig. 4(C). The currents re. #_ _ _j
[ q , sulting from these charge distnbutions are : N shown in Fig. 4(Db For these assump- C " ~ F,, i y tions the maximum current is about - 4o -zo o' .s ia e 4 o a i o o 4 e :s ' 11000 amperes. Por asovt a por IN lo 6 vo m qm g mgg
*6 The actual condition for rapid ad. *'"' 08 o ga m g 3 "5 L vances of the channel probably lies be- G'" CU" 'C" tween these two cases. They will be Fig. 4 Nature of charge and current distribution induced in the entension and the channel greatly enhanced by the characteristics !
during a rapid extension of the channel when the otal induced usese in the extension is draan that induence the transition from a glow from the channel benind the tip l 62n Wagner. Hileman- The Lightning Stroke-fl OcToven 1D6L = i ( o 90004024 "
1 ( i i
?
Fig. 5. Meawements cf remains the statement by Schonland that ( electric fields next to earth on!Y Iusec is reCuired to travel the length
-q y s v L. -,$W k k at pomts remote from the 5
of the step. This contradicts laboratorv h'
**W data which show that channels start at a low velocity and increase rapidly. Su 'C
[M.d.~Nkhh N:}E-UA~~?iU Md3P A- Rep'et of cathede. ray before drawing dednite conclusions, con-
?MI) csc.llosen that it!ustrates
'* " 8 "d * *f Fd' Sid'I II " "U1 be given to additional NMgagggynA*=%
e
]7.7".'"T@h b ~*
,,qg produ:ed by steps and by evidence. Th.is evidence is ducussed
- hMaure.m,vqdEM$%*inuan rm strete of Wu :omoewt cf a str:Le subsequently when the electric 6 elds, next
.Cr.;; d QZ{
to earth at points remote from the stroke. h*$L NWM. uy%::,=~.
~~ ~P'" dmile-*--
- x w w -.f:.~. n....;af % .,8 B-Actusi record that shows that as erth is apprached are eur.iu ,.
Q.a R Je - the field produced by the
, p. [gb P"w uy .- - M tee:s just .thead of the return PostTIvc Pot.antTV artes
, /."
fG - -M----EM The experiments of Park and Cones" v - m-w i streke become s M er and l
%p g, g6 - sg{ ,",y '7 -
ci st$ .r duratien C Jcob a a tvp c31 se:oon mamfest the same external character-istics for the formation of both poutive
$[U ..cu u$ k E~T.'c A r.w $-pj- d of the field re::rd tsi e' and negative space char;es. Therehre. i
{ddW':'::"ct.N[FdEMF.4C: ECDg 4~:'&-W:F;;, %q -- 1 hem Ae esrb f fe of the fe via assumin.; that steps rezult irotn the same f8) interplay of the space charges and the I channels, the negative channel formation *] should be similar in character to the i positive channel fonnation. Further- i more, since the channel formation and I
#q characterhties are essentially thermal
, a #
g k l
' <kf phenomena involving the absorption and the loss of energy, one would expect the ji I
i C) characteristics to be similar for both polarmes. f,l f Et rerMc Ftrto Mr ut'RDIFNTs NrxT j To Tut EaaTit ! resembling that above pomt .l. The in- m natural lightning because the two tense 6 eld ahead of the tip just mentione:t terminals may be separated by enormous in the past this information has ' probably supplies the explanation for the distances and the movement of the len.;th been analyzed in terms of the e!ectric , high velocity with which the leaders in of one sten would not make any appreci- couple of the charges cumprisin.; the g the steps of li.:htning advance. able char.ge in gradient from this con- strokesttM8 mre recently Wagtier" - sideration alone. Therefore, the induced has interpreted these measurements in f The currents described so far illustrate l the assutned phenomenon that the travel charge in the channel extension becomes terms of waves of char;;e and has concen- T time of the channel extension is 1,aec. an important element not only in ex- trated on those phases of concern to the 'l In considerin.; the case in which an open plaining the movement itself but in ex- transmission engineer. A dennite limita- j circuit was used at A of Fig 1 it is plaining the high velocities that may be tien of some of these data has been the s rewlution of the time sueeps and the time p evident that this assumption is incorrect. attained. for some current must certainly dow from But, while the photographic evidence constant of the measuring circuit Time j above this point. In considering the of Schonland supports some sort of ac- resolutions of I usechase been didicult to b other case current would cease below .1 tivity such as just outlined, there still attain. In recent work of this nature ; after a time of I usec, but would continue o to tiow above A because. after a current F 4 A B C D E wave is once established, it should :on. ,I tinue to tiow until its energy is absorbed 25 g sr ! by losses. Evidence to support the condi- l l
/o .-o j
' vEcToa CORCNA l g tion of Fig. 4 is offered by the photo. i 1 ,
W E% AL ; ST A G E 9u k graphs of Schonland, that show that nor f j ontv the ste;) and a short distance behinu
' {'I
' f %
OELD 37 ~cam j g tq f "- ;4
, o r 2 ,V S '
- t. }
it is highly luminous, but also that a dne .
#* ' d I e'
" " N' .0 A 1 M l l }2 5[
i trace of much fainter luminescence exists behind the tip of each step. This sug- k' / 5-bC -e:--- y, Q-r.[s,, j gests that current such as illustrated by -.g // p, i 3 ,
/2\ '38%
Fig. 4 is also being supplied from a long vE
,0 jE . VE7oR .
OH ANNEL I POTENri ALk o 5 length of the channel If the channel CHANNEL ST A G E 4
;j 'r 3
r development time is 5 asec instead of 1 / ((
;oacNA STAGE STAGE -i usec. the currents myolved would be correspondingly smaller.
{3-
-3 l In laboratory. produced sparks, as the l tips of the channels approach each other Fig. 6. Interplay of currect supplying the corona sheath with current in the chenr.el entenuon the tradient between them becomes larger of a step that leads to the development of an electric Held at the earth at a point remote from the and lart;er. This condition does not exist stroke of a character sindfar to measured values
*in 9 OcTonFR 10 tit Qrier. Hileman-Tise Lightning Stroke--Il 1'
90004025
( l4 I g , l Clarence and .Malan have utilized an Fig 6(A) the tip of the channel is shown step results in an electric field to which amplitler with a high amph6 cation and a just as the corona space char;e is de- the circuit of Kitagawa and Kobavashiu
; frequency re>ponse up to 3no ke. Kita- veloped tu its maximum puution and the can respond f,uthfully. and fur which gawa and Kobayashi" state that their channel is about to start its development the time resolution of their films is ade-recordmg equipment can measure a rise from a to c. All references to time are quate.
tune of about I usec and that the time referred to thisinstant as zero. Fig 6(B) The time interval during which the g resulution of the time sweep is less than shows the currents within the newly c hannel current has the higher currents 10 ' sec. developed channel and also for points is mull in ownparison w,th the total dura.
. Fig M A'. u hich is a replot by Kitacawa above a at thfferert ir.stants which start tion vi the step, perhaps only several
] and Kvbayashe of one of their records, small and increue with time. At "5 ucrwconds. But it does appear that j shows the ~1ectric field about 10 miles from pee the head of the channel reaches the duration is bnger than the I ssec
, a st roke. The field produced by the the b undary of the space charge and must referred to by Schonland. Perhaps the 1
f.rst return str' Le is indicated by the f'rst stop its ; regress. At this instant the sh cply rising portions of the turrent large sudden chance. the R change. The tharacter of the discharge chances sud- curves should be sharper than here in, small pi; e preceding the R chance are denly fro a thannel ntension ta a cerona duted and that 5 honland was able to stated by the authors to be about w d nchar ce A c >uuter .rdent bmlds up rewrd only a shat mtcr cal of the very j rec apart they appr%ch M then 1." incide the v rona sheath through the de- h uhest current mth the ph tWraphic rec durng the tin e just ;rece&g the vel:; rent of the space t h4rge .md the rmtivity mf the n!m that he ded. Per I R change .u.d are dt.e to the ' ten n the current he,;:ns to decrease. The manner haps alm, the resolution of Schunland's
, first leader of change is indicated in Fig. 6tC) at f'Im was not sumciently fast. .M ala n."
Fig MN is an osedlyram whkh was different instants. The solved lines in in describing the type of equipment he obtained by Kitacawa and Brvok? m New Fi.;s 6(Di and O t E) show the time varia- and Schonland used in South Africa, g Mexico. The nature of the fald measure. tian of the current at specine points. states of the time resolutmn due to the
; n ents preceSn t the R 1hans s +hown in A inuwledge of the vector p-tent:al of sc:ty of the lens as "p -ible tv meas-tz re detal Fit M C. in another repiot the current is : r sary to .ietermine the ure inter cals with an accuracy of a few d
of a very early ;trmn of Fig MBi by the elettric ,id at the ar und at a ; mint re. ma rr seconds " present authmrs and is intended ur.ly to mote it,m the str~ie. This is propor- The iminosity of the steps rh .ul1 give l
] indicate the eneral nature of the Celd tional to the mte.;ral of the ; ris!uct uf the sme i,!ea of the current carried thereby.
1 produced bv the s t eps. The ap"i current and the dmance throu h which it ' Incuer ally SchorLn P2 arrn d at a chances m H.;s l A, arm a b are f.ath- is operative linded by the distance to value of i6.uo0 amperes, but c"ntladed ful measurements but might be distorted to the point of iservatian. Since the dis- that the " luminosity of the step ; reess is
; sli htly by limitations in response. The tance to the N int of observation is sub- 'ar too weak to make it likely that the j droppin;; partion of the large R chan;;e, stantially cu stant for all of the points step carries a ;urrent of this ma.;nitude " l l is dennitely affected by the time constant in the tep, one need orly consider The liscussion just presented is not su:li-of the low cut d!ter used in the recording the area under the curves of Figs 6(B) ciently precise to warrant an accurate '
apparatus and should not be used to draw and 6(C). The dotted !ines in Figs. I estimate of the magnitude of the current. I condusions. Clarence and Malan* ob- 6(D) and c(E) represent these .;uantities It does, however, aTer a method sugest-sert ed sm:ilar Seld changes just preceding as a function of time. These are in;; how magnitudes in the order of sev-
. the return strcke and state that they con- also shown by the dotted lines in Fig. eral thousand amperes might occur. This l sist of a ' '. rain of steep and predominant!Y 6(F), '
is based upon decreasing the value of IL positive pulses following each uther at ~2-Except for the slow propagation of an 000 amperes, discussed in the development f to INsee intervals " avenge charge to replenish the charge of the channel formation, somewhat in i As a result of f' eld measurements in older neps and to su;miv part of the Sc h ,nlan d '8 preportion to a 5- or IUw see channel l was led to conclude s ery charge in the new step. the ratid changes formation perisd instead of the I ,.sec l carly that the current in the steps is less F in charge involve equal positive and nega- asumed at that point. than 10 e of the current in the main re- l tive values of charge. In other words, Further, it was awumed that the I turn stroke. Tlus conclusion was veri 6ed the press is largely one of rapid redis-by later work and is discussed m more charge and channel f irmation periods t-ibutions of charge rather than a transla- were about equal. The time of space
.l detail by WagnerM But aside from the tien cf a charge. Such redistributions do maanitude of the ground ,;radient meas- charge formation increases with gan spac- ,
not contnhute to the develupr ent of an in- but the time of channel formation is l urements, the shape of the records leads electric 6 eld at the observatiun point. in !cpendent of gap spacin::. T1us would
-g to important conclusions.
Thus, the only contributing factor ts the tend to make the time of thannel forma-j time rate of change of the vector potential tion smaller than the time of space charge
.M FCHANISM OF FROPAGATION of current. The time di!Terential of the formation. It is difficult to assign 3
; Fit 6 represents an eSort to show dotted hne of Fi; 6(F, is indicated by the denmte value for thc ratio because the cu.ditatively the ser;uential mechanism of solid 1 ne of Fig 6(FL The 6 eld at the time of channel formatisn also decrease
. the step. Fundamentally it is based observation point is proportim.21 to this as the overvoltage increases l upon the observed facts that for labora- quantity. The discontinuity anses from The detailed osedlupam m Fic 5(B) i tory ;;aps the formative current 8 of the the fact that the shpe of the vector po- clearly shows that the early stages of the g ace char,;e starts at a hi;;b value and tential thanges signs rapidly as the dis- downward leader are much more variable decreases st mewhat exp.,nentially with charge transforms from a channel to a than the later stages. Nst prior to the
- me and that the current 2 supplying the orona d!w harge It mn be en that this R < he ge, whHe ; cri A 'i the 9e;' ae
. hannel durmg its des elopment stage Schl is 'umiar i aerenl thoracu r to :he sull tiv er".iNe and, as uted Dv Kita-starts at a low value and inc eases almost dectne 6eids depicted m Fig 5(C;. *his gawa and Kobayashi. %ve re th< ed to 6 as a positive exponential with time. In interpretation ,f the phenomenon of the 10 usec, the magnitude of 6 eld changes are at f G2S Wagner, HUeman-Tiw Lightning Stwke-U Oc7ourn 1v6L N'
90004nn
t F G H A B C D E { I T:ME ? lN -5 0 4 8 13 C 17 u SCC 2 '
- l
! l
'y ; ; l , ; , a i o -
I
+
l , l l ', l ; > ;
'5 3 l
l l
. i i ,
i* d 15 ' I r I, , i ft z ;vo l - l , , , .
$ , l i f
$ ! s( ' '
O l l ,e , \. Y 1 l' \, , , ,
\, ,
\s \
/y {
,b \ l, s ,
;l ',< .
'l
. .r "
, : : e, , ,
i , ( / ! ! I g ', 1 / \ -
- t s /
s / ( , ( , s j g s ,
, /
- ' ' N .,. ( ,j , . , ,
- y. gl. . '
l vi, .ti a (TT: an (7 W .tiz ?A 1, s it : ~i < , i?71 iiii i < ,,a s i s i ti r iit a :i1 C777 ~iT G ,~ 1 Fis. 7. Stees in the development of en uped -henel j negligiNe in compar tsu u with the R the steps and current vartatmns in the ci the corona space char;e is entirely chan;;e. Nurinder and Stodregend also leader were so small that the light ap- independent of and a precurn - to the comment un the fact that in some of their peared as a continuous beam. The third subsequent channel development al- l, measurements of ground gradients "there possibdity is probably correct, partie. though the channel may be.;in before the { is a calm period before the be;; inning of ularly in view of the evidence a0rded space charge is fully developed. That j the main discharge? The step phe- by the ground gradient measurements. this work was conclusive probably can be j numenon is only an incidemal prehmmary attnbuted to three ctudiinc. (11 the , settin; the stai;e for the ;;reat incident - use of a 0 07X 100 psee was e that rose so f CtTRRENT IN CORON A MEMH rapidly that it could be re;;arded as a culminating in the main stroke. j The long perieds of the steps during It is possible to estimate the magni- rectangular wave; (2) the choice of a i the beginnmg of the leader formation may tude of current in the leader steps by'an sphere ir, stead of a sharp rod; and (t t! e f' be occasione 1 by the lin.itation, in the extrapi lation of the measured laboratory control of the free electrons and ions m the charge accumulating ability of that pr. currents. In reference 3 the current gap. In most cases tne voltage ruse to ; tion of the discharge within the cloud, or pip fcr a 6 foot rud-rod ,pp, for a po3i- full value and for some gap spacings the stated diferently, by limitations of the tive .mpulse vnltage of %0.000 volts air next to the ;;ap was stressed to as I leader's ability to mamtain its potential t9% of critical) was estimate 1 as M hi;h as 150,000 volts per em before a free as its capacitance is increawi with dxed amperes. Now apply these data to a electron or ion was positioned to trigger charge. As the length of the leader in- lightning stroke m which the voltage the gap. When a sl%cr :. ave that rose ; creases, the reservoir constitutin; the across the space char;e in its most ex- to crest in I see was used the results be-source of the charge for the next step is tended position is about 30,000,000 volts. came erratic and the crest of the entona ( greater and constitutes a more reliab!: To get an idea of the current that might be current pip wa. approximately propor- / voltage source. expected in a stroke to form the corona tional to the volta;;e at which triggering It is signiscant that so few photo- sheath at each step. simply prorate these occurred. It would be expected then, S[ i graphs are available showing steps in the voltages. This results in a current peak that with a copious supply of free elec- 3 latter stage of the leader development of 30,000,000AG0,000)(25, or SSO am- trons the current peak would be relatively just before striking the earth. Only two peres. Park and Cones" observed that low and spread out ove: considerable 3' or three such photogaphs are available the current pips vary over considerable time. This condition must have pre. and these indicate small steps. On the limits for a given condition. Taking this sented itself in the experiments conducted other hand, negative evidence is avail- fn.tur into consideration, and prorating by Saxe nd Meek and possibly others in able from the work of Berger, who has the data of Park and Cones (sphere plate which the current in a pointed electrode phntographed the last stages of the leader gap) and that of IIagenguth, hhlf, and over a dat plane failed to exhibit the - at a very cluse range of less then 1,000 Degnand (horunntal rod-rod gapi, it may magnitude of current peaks of Park and feet and with the exception of two photo- be concluded that the cur.ent in the steps Cones when the applied voltage rose to graphs has not been able to discern the of the strokes may be in the order of 1,000 crest in about 2 usec. presence of steps. Three explanations or 2.000 ampere,. Although these extrap- But the magnitude of the peak corona su.;gested themselves to Bergen (1) the olations are quite large, they indicate current is also dependent upon the ratio photographic sensitivity of his Alms was that the lightning stroke channel currents of the time to crest ni the applied voltage not suficiently great to penetrate the necessary to develop the corona are of a to the time required to develop the space intervening atmospheric conditions; (2) relatively modest value compared to the charge. For a lightning stroke step the P!exiglas drum (2 mm thick) through return stroke current length of 150 feet a time of about 3 to 50 which the light must pass may have The work of Park and Cones demon- psec is required to develop the corona absorbed the light from the steps, and (3) strates must clearly that the developrnent sheath ahead >f the channel tip. There- t l r Wagner, Hileman-Tire Ligianing Suoke--Il G23 b OcTout:n 1061 90004027
/, j A fore, a gradual rise in potential of the intended te consey an impression of the as shown in Fig. 7(C), the channel from point from which the corona emanates in, current flowing at the instant. It is as- both a and b has progressed to the points say, 5 usec would not have much effect -umed that the potential of the channel, indicated. In 7(Di further progress has upon the peak value of the corona cur. mdicated by the full line on the axis of been made and the current has ;;rown
, rent passing through this point, because the corona sheath. is .50304000 volts with signi6cantly At the same time the cut-during this time the back voltage de- respect to the earth. rent feeding the downward moving chan-veloped by the space charge would not Fig. 7( A) shows the position of the nel of the last step draws current from have sufIicient time to build up. channel at an instant where the tip, b, is the old towwurrent channel feedm; this 3tdl about 100 feet above the earth. It is last step from above. As was upl.uned Developrnent of Return Channel surmunded by its negative corona sheath. in the discussion of the step process the When even more remnte than this a corona prgagation of this current is sluwer than APPROACHING THE EARTH the speed of li.;ht; it is limited primarily dischar ;e belins to develop from the Fi? 7 is a simplified picture of the gen. tower and at ihn m tant has ainady de- by the qued with which the arc ;>ath can eral procenes that occur at successive veloped sizable proportions. From this accommodate itself to the higher con-instants as a stree strikes a tall yect position the thannel and its con na sheath ducurity and the speed with which it can such as a mast or a tramminion hne o,ntinue their progre.s to carth and when draw t e charge from the ' ace charge.
ic.w er. Evidence has indicated that as they attam the puition sh ,wn in Fig. Correpon.bn;: n6ers of the channd must the leader approches d. : earth the step. S N. m w hich the tip is 275 feet above utend intu the sp ace charge aW e the f become shorter in length and time. In the tower. a channel be;; ins to develop last step to collect the appropriate cut-I urder to discuss what occurs at and near from a. rent. And so the progress continues
] the earth, the steps will be assumed to be This distance is determined when the through (E) and (F), the instant at which ] ,
so small that the channel and its asso-ciated eurona sheath propagates with a average gradiett between a and b attains a value of about 6,000 volts per em, the contact is 6nally made between the main c hannels in the last step Fig. TC and constant veloc:ty of I foot per asec, which critical s alue at which brokdewn oc. I!!' show later instants at which the head is approximately what Schoritand ob- curs rhus ci the return channel has penet r.it ed served as the average velocity of the ,, ' y farther into the < pace charge and lowered downward leader. While specifie numer- = r. feet M a substantial portion of this char.;e to ci'000 x an # ical values are ascribed to the various carth. i. sta;;es, it should he realized that the This inctant will be desisnated as the Note both the prwress of the upward actual values vary uver the wide li:mts ref erence pomt iur time. Prior to this thannel of the last step and the time characteristic of lightning phenomenon. profuse corona had existed at a but the interval5 between the indicated pmitions
;.; The mast is indicated as being 100 feet channel had not begun to develop until which has e been chosen to suggest in-tall and the width of the heavv lines is this instant. Four microseconds later, creadngly higher and higher velocities of
,; the upward channel. As the upward R channel progresses, tentac!c like streamers d reach outward and upward and tend to
'i f spread the positive charge over a greater j j l
. area than that encumpassed by the chan-8 so ,- nel itself.
i, i g The current at the earth associated
' N
3 with the development of the corona sheath is small with respect to the current oc-50
- curring during the channel formation stage. This statement is based upon C ! j ,
laboratory tests. The surge impedance i ' l4 - ' of the stroke does not di:Ter greatly from
$ @d f the series resistance that is usually em.
g l ployed in high voltage laboratory test
' ) circuits, so that the fmal reference cur-x i lni e
, y r" rents should be proportional to the mit-s z j
..\
g ages in the two cases. Several ets of data, Park and Cones," Hagenguth, E
/. Rohlfs, and Degnan," and Was;ner and l j
$ l ,/ Hileman.' indicate magnitudes of the drst
" 20 " / current pip (current that supplies the
! [! //
/ g/ space charge) less than a few per unt
!i 90004028 ~ne- n- ~
'O i q N. [9fM ' [.I!k" and Cones show that when the crmcal voltage is only sli;;btly exceeded, esen
/'f /. F.ig. 8. Replot of the oscillo.
when the wave shape is ilmost rectang-i 9 s
/
/ 'C
.e p k' I grams of high streke currents uiar. the current trest is verv emall As e ope o the apphed was e os mes
,,co,ded by Be rg e r," trans.
0 3I" W . whkh -hould be the : e is the C 4 - - '0 lated with respect to time to 9 >2 6 2 pass through 40,000 amperes electrodes mas e toward rach otaer rl. wir, TIM E IN MICROSECONOS at the same instant the first current pip is even smaller '
-I ,
tl30 Wagner, Hil man-Tiw LigMning Stroke-U Ocimu 10til 14
l l l Both the rate of curret't rise dttring the so neutralized. This might be called
'F N. gfrcig E " T -
the induction factor. ph formative stage of the channel and the f;J 9" NW ~ fh Y terminal or final values of the current are Js R R a:Tected by the constants or impedances or. - -!ag tog - (to) M U '* I M the test circuitr, It is shown in the com.
.j' '
panion paper that the current increascs wheredQt is the charse in an elemental ring of radius et in the downward corona
-",.~ ,
O from zero along a concave upward curve
. T,k9 ' Q until it is litnited by the series resinance sheath, dq. i; the induced pwitive charze The Berger osediograms of the lightnin; in the upward enannel pr iduced by iGn strnke current show that similar force 3 r, is the radius vi the up.un! channel, cw Se act<ve in limittn; the Aw of cur- and R is the ra.!ius ,f a late :oncentru rent to an esantial!y constant vtlue after cyli:. der representin; eart h . The po- ;
fig. 9 Charinel formation in an anested tential of the upward channel i3 the same dischage :n a 6-foot red-ed 90 sho ing contact of the two channels has been established. These forces probably in- as that of the large cylinder representing th e outed fomation of c Al chwnds earth. The value of the neutralizin g volve such factors as the speed with which the low.currmt channel of the dow nward charge so cemputed is dependent up"n leader caa be conserted to a highly con the assumptions uaci For the values ducting hi fh current channel and the assumed by Waft.er and McCand the Cramt CITIES induced pmitive char;e was about 70%
. velocity with which the positive corona The mo3t extennve e s :.cas- of the negative charge. This sarne con.
head of the return streamer can propagate .i urements of stroke currents are those cept was later amphtled by R agner and made by Berzer" in Switzerland which withtn the negative space charge cyhnder that had been deposited by the head of the Hileman.1 Assttming R to be 500 feet, [1
#1 were collected on a _.0 meter steel tower downward leader in its progress towaid the radius of the are channel to be0 2 inch, ;,
ano, met mot.nted atop SIount -::an earth A substantially constant velocity and that the charge on the downward raivatore. The measurin g shunt was com na sheath was concentrated at a 8 att ined and some of the factors deter-locate I at the base of the 1% meter mast nurung this veh, city will be discussed next radius n equal to 200;:mor6 75 feet, they ,; j Berger has published 39 cathode rav rrived at a value et 0.42. The factor fj ostegrams taken from 1946 to 19[4 Induced Charges and Velocity of used is thus dependent upon the mech- pl of which the 14 that exceeded 40.000 anism by which the neutralization takes fi am cre are reproduced in reference 19. Upward Streamer pl ce. On further redection it appears yJ m tL.;r am3 B C H, I, J, 2nd K from agner ami McCann" showed that as that the channel probably extends up- Fj Fig *) of this reference are re$tted in the upward leader propaptes upward ward and outwarJ u indicated in Fig 7 Fi< % They were intent onant trans- within the cy!!ndrical voh.me charge This condition is dimtrated by the last lated with respect to time so a better deposited by the downward leader in its stazes of the arrested discharge shown _ idee of the extent to which ther co- pr gress e rthward by virtue of the in Fig 9, which shows the natt re of ordinate with each other wuH be seen. Usci'.lyram (B) is the most w ell-def,r.ed e!mrostatie relationship, a pmitive charge the development of the channels of a vertical Gfoot rod-rod gap for which the [} y is drawn upward frern the earth by induc-The others, with the exce; tion of iC1 for tion This charze may be viewed as volta;e has been chopped by a parallel , which the zero of time was clearly in- gap just prior to the union of the chan-neutralizing the ne;ative charge of the p dicated, were drawn so that they coin. downward leader. In the Appendix cf ne:s extendinif rom both e!ectrodes This g eide 1 with tB) around J0.0m ampetes reference 23 the foilwing relation is de type of mechanism indicates that the ra- i Thev all indicate a variation with time rived expregir.; the proportion of char;c dius of the upward channe! %, of equation tha[ follows a somewhat posnive expo- h nential curve Other available oscillo- l grams of stroke currents are reproduced in Fig 3 of reference 1. With the excep- l 1 tion of the record obtained by means of a captise balloon, which is not character. 7 d istic of direct ~round strokes thev all r
-, e, : Facu stastas exhthit this same upw ard conca ,e char- ! gg gg q 2 s r q acteri3 tic.
As sh m n in the compankn parer all of d f 3,3 h / n the zap currents for rod rod gaps exhibit j $$ $ */ i
~ '*
the same positive exponential character- soo4 o iz o ,- ./- - yg ,cc r y istic with time It also d. lustrates that ; !, y!
. the time lag of rod-rod gaps is the same sco l 0 o Czo- of -rA 7 / .
*E%as pa % j /
expressed in microseconds for all gap j $ spacinc and is only dependent on the *@j 008 1 'x // fractional amount b. '.ich the critical j,f N -
/ /
voltap is exceeded The times to crest b8* # W 'M"
- l. / / t N' l of the currents from Fig. 3 do not di:Ter 2co ) 0 04J -
curve computed to have a e CF greatly from the laboratory curves, de- terminal velocity of 0.12 e j ,'!-
~ %g.'
- - L,.-,' [, /
spite the great differ nee in gap iengths, and to be proportional to 11_ the curve for comparwn is one in the eunent eurve, Fig. 8(B). o I' o e 2 is 6 The conesponding cis'ance o 2 4 s io which the critical voltage is exceeded only curve is dso sho.n nut N m st:0Nos sin:htly d31 CCTohER 1961 Wyw, Hilcman-The f.igianing Suoke-(( , 90004029
]./., ,g o
i ,
- The following simple relation then exists height is approximately 50 meters. The l 10, is much larger than previously thought 4
between the charge per unit length and lengths of the last few steps before on-e- and that an induction factor approaching tact with the upward channel are c!early the current: ; p unit might well be achieved. shuwn. Whether the actual mechanism of #
) 1 neutralization is of this character, in- 0%
? p
- volving a rapid movement of positive If, on the other hand, the charge at any Tontt:ostry or PATH f charges into a region adjacent to the nega- section transfers inductively to the down- , S Schonland observed that successive tive char.;es, er v Lether the negative ward channel at a cunstant linear rate as a charges also move in toward the positive function of time, with the velocity re- "fES do not genera!Iy foHuw the same
{ direction. The tendency to maintain a
; channels is not known. It is probably inaining constant, the current also in- ,
similar to that which occurs in the last Itraight h,ne m the progress et the channel
; creases at a linear rate until the crest is is not strong. Since the channel of each l stages of the laboraturv produced spark attained. Thereafter, the same simple
~ step starts from the center os high charge i where hi.;h terminal velocities art pro- relation cO,ts bem een the current and the e ncentration, the direction taken by the duced. linear charge Jensitv. If the head of the cha:mel would be highly Hnsitive to A number of experimenters 2 "" have rcturn .hamal mUves upwa d with a 31 ght differences in density. i aere may L measured the progress of the tip of the velocity that is not cunstant, but increases " #"I 4 channel as it moves across a laboratory prepe$sively with time. the abuve <imple b* *"U ""*but directions [* enty "3 *rmena h'"*lly dvmina gap and from these measurements have relation between current and s harge e dominatmg me develops a luw sr :m- ~
< computed the velocity. They all show den:ity continues to hold if the wave of pedance, and in the process of forgmg that the channel starts at a low velocity charge remains truly rectangular. This ahead decreases the electric 6 eld parallel
'q and increases rapidly with time. The conditicn also requires that the wave of to its path m its vicuuty, and also the result resembles the shape of the current- current remains rectangular, that is, that tangential components of Geld of the time curves. it increases equally for all points bel.ind I ;gm.g channels so that less char.;e sep-j Schonland, 3f alan, and CoIIens, dis- the head of the wave,if it is accepted that
' the velocity of the heat of the current "T"lI"" Is requ red m, Gem to neuttake i cussmg the vek, cities of the return the Geld .m which they are advancin:;.
j streamers oflightning, state that the range wave is shnilar in shape to the current-y of variation is from 2.0 to 14.4X10' cm/ time wave- No accurate photographs of see and the mean 5.2X10' em/see, while I j , the first few hundred feet of travel of the General Discussion I the value :ndicated as most frequent is return channel are available frcm which 3 5 X 10' em/sec or 0.12 c. the velocity can be estimated. If it is In laboratory gaps the space sbrge fvrms at a substantially constant velocity In the companion paper evidence is as<umed that the f'at portions of the shewn to indicate that the channel cur
- strcle currents, obtained by Ferger," and thus the time of" formation is pro-1 rents in rod-rod gaps bear the following arise from the induced charges rising up portional to the ;;ap length. On the other hand, with gap tengths of at ! cast j relation to the ,mstantaneous velocity: the column, after contact is made be-tween the - ward and downward chan- 100 inches as shown in Fig.10 of reference i m 2xto- y (zu 3, the time to breakdown is suMtant: ally nels, then it can be expected that the where i, is the current in amperes and V terminal velueity of the upward channel independent of gap iength. Thse rela-is the veh city in em per sec. If the most on contact must be (using the most fre- tions indicate that the time of epace frequent return vehicity of 0.12 e is in- quent value) in the order of 12*o that of charge formation plays a mure important
'4y serted m this equatmn then ny becomes ligh t. Fig. S(B), which is the most clearly part as the gap length increases and 11,500 amperes, which is gratifymg!y clos' de6ned of Berger's stroke currents is re- emphasize the role of the charge forma-j to the most frequent magt.itude of stroke plotted in Fig.10. Assumin; that the tion in the step mechanism of lightning.
current of 14.000 amperes. vekcity is proportional to this curve and The forcing electric 6 eld at the earth is Or, viewed differently, if a rectangular fixing 'the terminal vehicity at 0.12 c a applied s'uwly in steps as the dt anward wave of charge moves vertically from the velocity curve shown by the dotted line is leader a;;proaches and then more rapidiv earth at a constant vek, city, tc, then the obtained. IntegratinE this curve with as the tip of the last downward channel associated current is also a rectangular time results in the dot. dash curve that develops. In this way the positive charge j
- wave. If the return channel can be repre- gives the travel of the head. The dis- at the earth has a somewhat lon::er time sented by such a wave, the implication is tance traveled in the time required to at- to develop than the space char [e ahead l l
that the head of the wave on reaching any tain constant velocity indicates the point of the downward channel as the Ltter can vertical sectic,n instantly annuls the por- develup intensively only after the thannel f above ground at which urnon with the tion of negative charge at that elevation- downward channel occurs. This distance, has completed another ste;> to a new posi-l about 400 feet, falls within the range of tion. Or if the head of the duwnward expectancy. channel moves at substantially cunstant velocity, as the ground gradient rneasure-
' 5~r"to y
~
5" " Ev:r ESCr. OF I."PWAkD LEADERS [ ~j others mmcate, then the sse . a.,e :s l Fig.11 is a photog *aphic reproduction continually being formed imm an elec- l l
?
! of a sketch showing the kwer end of a trode that moves at the average efoc.ty Boys' camera record obtained by Dr D. of about I font per sec. The < ace j
}
J. 3! alan and presented in a discussion of charge from the earth terminal is being l upward leaders by Dr. R. !L G41e..a formed from a stati nary , " le. The I The nint x indicato ahcre the downward current during it.is nter -d ap :n; ie leader terminated and was net by the charge in the wr..na sheath n .t ively I Fig.11. Boys' camera photograph of suo'.e small. The signi6 cant current only oc-to ground upward channel from the earth. Its
]
1i%gner, fli!cman-Tlse 1.igi:tning Stroke-II nc %. G32 x 90004030,wn1961
E curs after the head of de upward channel type assumed here but, whatever it is, about 0.001 c, the velocity of the dart j has attained a high velocity, such as in the average 6 eld might constitute a leader is about 0.03 c. As the dart .,. exc- of 0 C c. mea <ure of the conditions necessary for leader appr saches the earth. ther nmt The ;- of the steps in the down- channel deve! >pment. The space charge still exi.t next to the erh 2 sin ' - w ar l 'e i t lw:1 wt be acey,ted as mu be relate.! to the total potential ch. cl in hich tb san e< H . n i evidente o.ar the spate char;e formation thereby produced asit is the quantit y e,Mse exi< t as in the channel thr' u;h is com; im c halted at any point as might dere!?pment retards the downward prr ;- the dart had develved. Theref t' a be the ca-e when a dednite irnpulse voit- ress of a step. Also. it was seen to be a similr dart lead r shoulI sprin; tr a the age ~ anhe ! 'o a cy!!n !ncal wnduc mr meas .>i the critical sparkwcr ;ra !!ent. earth and meet the d unward notinglart . and th. exe charce de.d ts to a Rr this purpue an average .;r aS. n t le a ' - If it is true. as wi"'
- l de ry i RaA - it :h ul l ' u:' te as d wit; e i 4 :P ua ve: d ; 1, th u d. em nt tne th%ht i e a centinuin; furmatien, and connant gradie:.t. The oth. rde of the with the ve! . city of apprr ach of the when 'N conditions at the ter ninal are charge distributed within the downward channels then the current should have a fav-rati, * . the irmatian of a channel leader the rnanner in which it odects stee;er sb;e than the tirst comp nent the ; m; - of the channel is renewed. the potential of the conduct ng channel In Fi;.10. for exam;'e. the e :rn nt sh uM The < <r v!iti . , merely mer;e into compriein; 'he core. But this i, onb aa ten I to 3 tart m t he vaine at a hkh *h-ew ch - A smiLr p: en. men .n is fact r that 2:Tet , the pote a! The ve!ocity n m and the ts era; 2 .i j
- re c.
- C.- k W . i ? kes sub-eiuent char.; in othe porti; , itheleade an1 rice sh< uH F e hisher. The latter o to the - < T htee '.ea ler, had bee the di3tance from the earth are aL t' vi the current curs e shou: . be r -
tnou.;ht to be continuous and propagate portan t. The neld outside as we3 as in- in slope to the latter stage > i the current - with a vel" city vi about 0 01 c. But the side the space charge determir:es the of the first component. b i e!eetric de!d measure nents.' at points potential of tbe core. Thus considerable Recently Ben;er" presented informa. ! ren ot- ir m the stroke, have dem i depar'ure from the 2numed char;e dis- tion concernin ; the wam H pe of the i i strated that they mv. lve fa,t rep e2 tin; tnbutan might be pe mide with. ut curre '- in cmnp ne 's <u b- ' < to pr 4 . -c, . .t h t pen ..! ahi .ut 10 gee greatly adectin; the ; otentif the tirst, that were obtamei ..n % unt San The irdinati"n between the various : alvatt re. One chart shows the current , cut nem of the stroke can be illus- in an 11. component mke. The tirst <p Multiple Strokes tratss.v a num ncal example. Consider compo ne n t had the usual euncas e-u p- ( a r e c: r sachin'- the earth with a . ward shape and rose to 30.0W ampercs i The fact that multtr c li;u..tning str%es 9 -) cec The cren 'f *E cubqt:nt s, 4 , e ,'he earth t W.. .. . take the same ' oath to ground m. as ta2t ort . its Auume that the average co mpone n t s were smM:et but the rise t w' the efects of the precedin; cornponent gra,la : . t the sp ue char;e 9 tween the time was leu than I rec. In conr. <a-have not become :.uuy d2supated by the head of the duwnward channel and the a tion with one of the authors Bc.;er time the charge.;atherin; mec..anism $i , ear th nu esary to mitiate the e. nann-1 within .he cloud had attaired su:r aciant stated that this ininrmati >n was h.eed p-f r. . 9 carrh is Moo vu!ts per em. As by the tri; rin*' time if the cat h ,de ray - potent:al to brea.,: dw n the wu,.ent.1 n can Se -een from equat mn 9 this path. The usua, explanaticn given f.' osc@grat h and that the %rtien of the 'e i the he:.g ht ut. the d mnward record prior to this was lost. These rate, dete'
, the -ub3equent strvkes vekin; the same [ l ch.m i at thi-inent as 1a teet F r" hh .
- ion zatica in of rise are much faster than for the drit 1 ,
TaNe I m l Fi; it can be estimated component and may be signirbnt in r. that 'he charge distnbution a1ing the h M h mmt art ci 7 directice edeet stiU reains 3.!cCann considerations adecting tran4 minion !ine 5 str. + -tnt distance back : rom the tip is g gg g g ,x peric rmane. Both 'he magnitudes of l ah,ut iu ; 0 -' coubmbs per em for a explanation by stating that the nain the subse<iuent comp onents, and the.. str & P tential ci 50.000MU volts return stroke current heats a sma:1 tirnes to half due, were less than the 'irst - Now ;f tt a usumel that the velety at the had ;f the upward channel at and d- d s m a M- @ Wera. co m p .ne n t. This information is still too j ' y,d i ion of the stroke. Te34ef to df 2W stE"f1I COCCI"SioC5 C" af t er c. .atact is 0 :2 < then the streke the ,neated column d.m..uses into the sur- f current at which it ievels od :. . from equa- . . roundmg air, but as it dnes sa it forms a
""" I2 column with a larger and larger radius. Comparison With Other Theories i , o u , , ( p m 5 < 10 " whose density gradually appr saches m r. {,
mal . Since the breakduvn voltay de. In a subject as com E:ated as lightning y
"#I* creases with decreasing density the pre dNhar.;e it is di Ecult to determme w ho and ti the setv s 0 l c. the strae ceding path oders an cader path far the originated certain emnp.nent v perts -i current .s n mo amperes subsequent discharge the theory The novelty of a compte ,,
In the the rv pr p ed here, it was But regardless of the explanation. it theort resides rather m hnw the com. p asumed that the charze den.-ity vaned in- is an observed fact that the velocit. of ;;onent charr.cteristics are asembled verw!y mn the raaius which results m a propa;ation of the downward leader af an i adett each other. Little attention g c<nstant electric gradient However, components of a stroke sub5equent to the has been given to the behavior of the $ this mumption is not critical It should drst travel with a much higher velocity stroke just before reaching the earth I be reciled *. hat one mk of this partte- than the Srst compc nent. These are and the mechanism of the return <trae ular quanttty is to indicate when the called dart leaders. They are almost free Without attempting to cover the ennre combmat e s of deld and charge derst:y ci steps. a!! hough Kitagawa and Ka- neld, several theories that have receis ed p reache3 sush a critical value next to the bayashi' bave seen indicat:ons of very widespread attentien will be desented arre-ted channel that the channel wdl high fre<1uency steps in these discharges brie:!y. agam resume its progress It mignt be also. Whde the average ve!acity cf Schoniand" c<mt!uces that the char.;- that the actual btnbutim is not of the propagatien of the 'irst comp"ce n t is advances downward as a cylindncat, Oc tm k M1 ll'q n e r. Hi!mta n--De Liftiniq SuAc--fl G 90004031
weakly ionized, body whose surface a gradient withm the pc. ;narge. The face of the envelope, and the other t next rtep develops when thrnuch chance, theories view the phenomenon as more of gradient is 30wo volts per em or some-a protuberance is forrred oa the enrelope the nature of a tric;ering phenomenon. i what less to account for the lower air d<nsities at high altitudes. This pilot of the bulbous spue charge that may The long periods of rather uniform steps leader advances at a emstant velocity. occur anywhere on its surface. This pro- would seem to favor the tripring view-It is accompanied in steps by a highly duces an unstable conditior, and, the new puint. g ionized leader. The highly ionized leader protuberance is enlarged, it draw, its As was mentinned at the beginning ed cannot advance into the uncharged space charge by means of a plasmr. that pws this paper, the real Uterest of the authors ahead of the weakly ionized pilot. Sup. inward into the last bulbous tip. .ies in studying the sequence of events as j pose une considered an instant shortly The theory presented here is a cum- th <troke strikes the earth. Aside from 4 the work of Bruce and G'.lde " Griccom.
; after the leader nas cau;;ht up to the filot- bination of the<e theories. Cetain aspects They advance tqether in a field of about and ti.e authors,little has been done al"ng are drawn from each and certain addi-j 3M00 volts per cm. But as the pilot tiens have been made. The small- this tine G"ldeu asumes that as the ; advances the necative charge produces dbmeter < ,nducting core is omi!ar to Lwn xard leader apprwhes the earth.
a negative Jadient behmd it that is that pren ised by Komelkov, Griem. corem Mreamen derdup when a gradi-directed in e opW te i direction to that and the authors in a previous paper ent s a%u w wo volts per c m a .t - ri the pre pe!!in; Ceid and eventually the The shape of the < rona cheath is e , ne- tained. The reducti..n frem 30*O volts net f eld drops behw 6.NO volts per em what like Komelkov's. All the theories pe r em is taken to ; rovide wr mmor
! at a point behind it. At this point the accept that the gradient in the eurona irregularities such as grass, shrubs, etc.
advance of the leader ceases. This space charge is in the order of several This he takes as the condition for deter-rapidly produces a positive space charge thousand volts per cm, but only Schon- mining where a stroke will strike. The 1 ahead of the arrested leader because land recognizes that the averare value authors believe that a corona sheath will develop much as G 1de propmes They
! electrons cannot be furnished to re* must exceed a certain critical value to plenish the space charge. In the mean* encourage the devek pment of the exten- do not feel, howes er, that the exi-tence "I time, the piiut advantes by virtue of its sion of the ;1asma channel. Rwever, the enri na 4 cath at the earth is the < on.
self-inductive property. With the de- Schenland is omewhat vague as to the trailing enterion. They feel that the y s ei.$ ment of the positive space charge, a actual esistence of such a plasma channel signincant facter is the instant at which gradient of 30B00 volts per em is again within the pih>t leader. It is I robable the au rue gradient between the tip and quickly e<tauished and the thermally that the actual trancition from the gl .w a oint "t t he ort h. of ahnut DH < tr ionized channel Klin advances. to the are is ci the nature pr< pmed by . To s < lts per cm. is attained. At this Komelkov," " on the other hand, as- Bruce but on a 61amentary basis. Thus. p,. int a e nducting pluma is for:ned from sumes as do the present authers that the the space tharge in advance of the the point on the earth w hich menes up-channd tip is fed by a thermally ionized arrested channel is fed by a host of sla- ward with accelerating velocity t~ ward
; channel having a drop of about 60 volts ments or streamers and the current at the the slowly moving downward conducting per em and that a corona space charge base of each one increaces as its length in* plasma of the leader. Griscom's pre-hasing a gradient of between 6,000 and creases. At mme instant, current in one strike theory proposes that before the tom 0 ~1ts per em is produced ahead of of these 61 aments reaches the tritical develgment of the currents of the return
, and around it. The stoppage of the value of I ampere and at this instant. an <treamer a pre <trike current, from Ahich i !eader, according to Kumelkov, is att-ib- are plasma begins to i; row from this point- this theorv derived its name, occurs at j uted to a decrease in the potential of the This may occur in e.eral %ments al- the earth 'whe+e magnitude is about ' hat I thannel tip caused by the need to charge most simultaneously. But ene of the of the return artke current and whose l-up the new cction- Slaments eventually p revails in its growth duration is al>out I cec. During this Druce
- atinbutes be 3tep formation to and because of its shielding effect upon miermcond, the current ; resumably rises the rapid transition from a glow discharge the others robs them and eventuall) from zero to the crest value and rMarns of high grauient to an arc discharge oilow emerges 2s the new channel The direc- anin to apprmimately zero bef~re the gradient when the low discharge current tion taken is determined by chance which current esociated with the return etroke reaches the critical value of I ampere. would explain the forked character of the occurs.
He also postulates that as the are has stroke path. This is similar to the ex-paused and starts . n its new ,tep a lateral planation of Komelkuv. Improvernent of Line Performance fkw of current to cupply the glow occurs- No doubt a temporary drop in po-The carrent at the tip increases as the tential of the tip of the channel as the new While the purpose of this paper is to length of the step increases untilit reaches step is formed exists just as premised by analyze the lightning stroke ( harm ter-appruximately I ampere. Kumelkov. But the authors feel that istics with the aid of the available know!- Criscom" has proposed another theory this has a secondary rather than a primary edge concetning the predischarge current with a highly conducting channel that infuence The prennt theory dirTers of laboratory gaps, a knowledee i 'he 4eds the tip. This theory is radically from that of Griscom but agrees with that htter su c ests a m. idinca t rn f :he different from the others. His concept of the other three in the assumption that manner in w hic h the impube har scier-is that as the discharge adcances a great the extension of the conductmg plasma istics of insulati..n are applied in he o m bulbous volume of charge is forrned at the begins from the tip of the tast channel and putation of the lightning charactsrecs end of the conducting channel many grows toward the earth, rather than from oi lines. It is not ,enerally appree:ated times arester in diameter than the some spot on the peri; nerv of the c'rona how inmortant the y redis h.. . curwnts e- wirical envei te of charge behind the ; heath hack to the . " nductin., : mnel. : < r' d n t' pes "i ga: 31 k tip. As the charge volume grows. it The Criseem prestrike theory h n the Fig 12 hews a gr. nd airt c ' o Yd mamtains a surface gradient of 30.000 timing of the inception of the new >tep as to the top of a tower and a est volts per em, although mention is made cf a chance distortion occurring on 'he sur- suspended from a town arm by means of 034 Waver, Hilcman-l'he Lid:tning Svokc-Il Mma 1%1 s 9000L 032-
en inutlator string It is assumed that eficets can be obtained. Consider this from the cloud to the earth. The currents ' in the steps are small in comparison with the tower top, tower arm, and ground same insulator string. Suppose that a the currents m the return stroke wire are all at the 3ame potential w.it h triangular pipe. pipe gap such as described
- 3. Frurn independent meaau r emen t s of re'j a i ' to ground w hich will be desi;- '
for Fh;s.19 and 20 of reference 3 is u<ed. , electric ndds nest to the earth it p. ,in u n a o 1 ,a c , . Ale. For a breakdown at 4 pec a chargm.; g. , g f n,m ex ek h h 6 i 3 ,m volta;c of abi>ut Ma is ret uired Mr the g th it thu rt e ney vf the i npt!, ; pr ,. .u d b> the stepI he nes ., man r u the earth e, . po itt d of conductor at the tower 6-foot a. The predischar;e cu rre n t r, = eurrent do4ng in groun 1 wire from is about 600 amperes and far the v. Itage is approached; they dimini+h from ab.,ut em.31 "a., 50 to 100 Mec at the clou i to about 13 see i = curqent SwiN in conduct"r from each acras the gap is about IMURom volts near the earth. One might soume that side Camparin; Fic 14 and 20 of referera.e e. e ste 1 wh m orm p. :sp re.
.-s -s ,
,, t of gr. umi w ire a "",
- 3 it wil be 3een that the ;urrent is duce ! TM 1v! action is rei d sreit bv e, e A t . % the sr * - W4 M L i P'i"Ib
- M h O" W ' u s g o.r & C " o,.@C 7,, , .ml surge impeda'tu betwe m 5** '
conductor and gr >und wire same overvoltage conditions. So, if the the earth. 'rt lenW jun pnw to remmg gap is inere tied to 12 teet the terminal Then by anlying conventional wave voltage will be about 2,000.000 volts " *PP' **"""' *U'
- sumed that the downward leader, as it 5 the .ry equations which matchc3 the in3a!ator st rity w,h-appr.> tche, the are moves ai h 2 c. n , tant
;13i mer volta.y 2nd the grelishar;e eu' veLcit y w it h. m ste, WL 2 cer tu, g e.g 4 point in its pr w is re.rhed the eic:'ne rent will be abi>ut 1,2'10 amperei The c: e 2,:.-2n (14- IT ' d i"' t "" "hject s pr7 N fro the
% thmu@ h lim i@ra , then ,
earth excuds a v .n u e ai .m m on, prr Eliminating i, from thue equations ' O M " MW L2"O ' 2 "r J W oO Nts. cm and corona stromers bedn to form This amounts to la (Fe of the insulator at the earth. These streamers euntinue f
, 2,5 ,2'-2 J string voltage to grow and a corona sheath is formed
, about these ob jec t s. As such they are i 2 2 not im por* an t but the space charc: asso- 1 and lettin.; the coupling factor 2.,i '2 be Conclusions ciated with them has m equa! inn; e&ct desigwed Sv K upun the tieH betuten the adsancmg The authors have attempted in the ch:mnel and the eart h . The current at ce = Kq m l - K2 Zie (161 the earth associated with the formation of l light of laboratory data to determine the corona sheath is small with respect to ,
The vi itage acrosa the insulator string e, what happens durm.; the very compl.icated the current in the return str ok e . Only 1 phenomenon of the lightnin:; stroke. Be- when the advancing channel has reached 1 I is cause oi the modest limitations .,f 'abora. point where the gradient bet w een the [
<y%< .In 11-K'21, 19 chanm ! tip md 'he m A or profu y ;
~1 tory f aelities it is necessary to extrapobte obxets redes an average value of 69 0
= ;l - K q -(1- K2 Zi/2 (18, a great deal to reach the grand scale et. yg, m m due, an are plasrna form from natural lightning The following is a th- object upon the earth. Prior to this ,
brief summary of the theery advanced in im at the enttre tower top of a transmis- d where i is the predischar.;e current taken sit ., line tow er might be en3hrouded in F bv the inndator string. this paper. co: sn i w ithout serious con 3equences . T,ne n. ret term is the component of volt- 1. The downw ar i leader uf the tirst Theicaiter, the current at the earth in. l age that is u3ually considered in line componen t of the stroke consists of a creases at an es er increrung rate forming i central highly conducting are pla+m a a concas e upward curve with time until I calcuhti<.ns, but the actual voltage acros; the iamlator string is less than this bv channel of abirut 2 mm in diameter which the maximurn value is attained. At this has a drop of 2 bout 60 volts per em. The poin t the channel tends to datten utT P the dr. p throu:;h the line impedance (1- conducting are plasma is surrounded bv 2 ' h. KV 2 This relation is true whether negative space charge whee radial gradient d is ahi.ut 9.too volta per em. The radius 0, the computed value (1-K)e,, ne ;!ceting j. is apprmimately 2N0 feet and .lepends 'Q '9 the prediwharge currents, is produced by ~ ~ upon the potenti 1 of the stroke the stroke currents in the tower and ground wires or by electrostatic induction 2 The steps of the first component are j, I"**d DY th" 'h" ** d"2" d N: from the char,e ' in the stroke above the central channel and :ts mg surround.'" corona 'I Ch' // b
- tow er sheath. If an nstant at which the central eg [/f '
l, In reality, therefore, the significant channel has just reached a point of masi-voltage that should be used with the con. mum progress in one of the steps is con-sidered, the corona , heath advances he> ond g [p'
- p. ic 3, ve n ti, ,n alle compu ted values for the ic 1 his pomt in front of the channel tip. ]-
in3ulator string is the lab <,ratory obtaine ! ',R ea s. m a hie a greemen t ts obtaine l bv j volta.;e acre the insulator string or other ex t r a pola t e g, from laborator y data, t h'e ec l form of ins dation plus 0 -lGZi 2. time of formation of corona pubes of g2ps , Tests have no been made on a bare string to the tune or formation of the tnerement of the corona sheath When in the progres+ A of insulaters but su ppo*e that for 15 of the des elopment of the eurona sneath, a .. U.Vinch insulators for a bsee time la? the energy concentratmn at the tip of the ' the dashover voltage is about 2.000.000 arrested channel reaches a critical value volts. then if the predischarge current is or possibly when individual corona stream- - q es reach a certain entical current, the about 300 amperes. Z = *>00 ohms and progre3s of the channel is again resumed A, "' O I The high speed of the prr>gress of the channel step is exp'amtd by the high charges I / il - K "24 2 = M.mo volts (10 induced in its tip, because it constitutes *p an extending pencil of high conductivity 1 or 3.9ea of the dashover voltage that pierce, the retatively stable space Fig.12. Diagrammatic sketch of tower with But if steps are taken to amplify the chrge .,f the coron2 3 heath. This phenom- 9tound wire and conductor showing nomen- ,, clature used predischar .;e cu r re n ts. much greater enon is repeate i as the :cader head tras els
%gner, Hilmian-The Lightning Strue-~ll dM L .g OcTou n lodi 90004033 0
4 .
- 25. Ocevaeuwes or t'rw a no Staesuses su The velocity of the tip of the downward 4. Psoonsnavn Lic urvrmo-!!. B. F. J. Schon. Lio n t wt w D8scu *ucas. R. H. Golda. .Vas ca.
l*
- d. D- 1 M*l**' ondon. F- C ' "2- P "#'"rs. London. E ngland, vol 160, sept. 20.1947. p 395 channel is approsimatelv 1/10 of l'~o that Royal Society of L
'"Londc s, '"E ngland.
I of light or 1 foot per v'ec so that it m:ght series A. vet 152,1935. pp 595-c:5. o d N***'"""C"'"'C""'"c" ' be viewed as an electrale at a constant P*s vui na Di g n 4 nc.sf =fe atwLiku
~^
4f {g' = Evof [2
. n. *1i3 pp potential of several tens of millions of p , ,.
52 , volts progressing toward a flat opposing Maan. Curseri.v Jo rnd. Royal Metwr*gice p.uum, by K. Derser of R r.ron t un tus I society. London. En,;ians, vot 83.no no. Apr. :7. electrode or a projecting rod. As this W57, pp M 1 -71. woes or rein Is tr a s4 f row at Strov Cow ui rr6e uccurs the gap length is decreased slowly. " '"***'"" 8 'N 'd " H ' 8*
- t 8 '
L '"".""J f 4. C l G R E ,"%r,0. Thus, in referring to Iaboratory data fof 6 F t rt > C is 4 sots AND V 4 N E A nWS OF b V WI- Tw;lr: q m sontry Den to LMurs No Ft mn N. Gnawa, **** P"*"' '~""'#" guidance concerning the phe9mruenon, one # E'" k Me rn y M Kota5 *5hi- M H ^ w c' a."V. S. K om e'is or." h.0.t ha. j nust look to the current that results when t "R
%nio, ece nt Per.~smon Ad'
- ace' Press. Inc'o ^'
Newm".'.nc M- NR. Tm 'ac al hn i 3.iectrmty ' as pp i sts,01. of me . an im;.uhe , s uc age that will just cause or g , s. V., rection. no . 6. I K rp 9 *4-* 4 breakdown is app'.ied to a gap. In these 7. A Cears uisow or I.s tr e,ctvvo axo Coveu- t<* Tns Iwm uss t vr Et scipic reen 4 =un j cuoen ucn tmo on.cn oces. N. Kna ga wa. 3 cases it is found that the current increases 19 Low G Ars. V. S. Komelkot IM. no ( at an tn er-thciering rate and extrapr.lating M. g.oot J w .i g g Gm Wi.4 Fe m 4. "'M iro:n IJbor.itory din Mid'ns One finds an 7.'ehiven. D C . ml 'i4.1%O ?? I I 5'I ~ I 2O I-j y p, y g%g, 30 Ta n 1 %:rl A r!OM br Lv.%G EO R ! klC AL D L S. f acceptaMe agit+ ent in the durati0n of the g g ,g cn om C. E. R. Zh u,c e. Ir sto l 5 .4.hy a. rieng p ,r t ion of the current. .\ g ain. H. Cottens. Trxtst'. N yal F u rty of Ltn.I n n . of twivn. vol 1 s t l'J 6 4 pp _.
.ac t >
! vol IC WJ L , p C44L I argumg fr .m lh rab.ry results w hich ' ' " " " ' " ' ' "
is r u n t a .p r st .c. - u z. 5. B Gr acom. show a corr %p miing increase in velocity 9 Pe m rmv s Liuhr%INo !!!-T n s FWs 4/TC l ? rat citw b .e ins F r.tru s 5 ap s. D J ,ja,an, 3 ;,, , g;g g g,, _gy,y , f 3 .n ,, , g , a I the tip as the i. .rrent nerca es, one H Coi!"D8- t h' d ""' 1 ' 13L PP " U3 a~N e n A uJ3.
' night evnc!uJe that a c. trespon ling in. 3: Tas t :a rmo h n4nos C. E. R Er uc e, Create in velocity of the ret tirn channel to Peucatswys Liourwtho '.% A Car terso" R. H. Goid e. br u a', Imt a ution of E cetncal also occurs. The limiting value of the or Puorocaarnic awo EtsciancAL $ rctilEs Or E ngiorers. London. E n gla nd . vol 48. pt. 111 )
ran DMCH A RGs P aoc ess. B. F. J. Schon'and. v ur E=p.unnen no. E Dee M, n em j
' selecity is that of the observed head uf the D. B. Hodga H. Couns. NL. mi MN brightly luminous return stroke which is p.56, 33 Tus F am Ehey or Occt mas ses Aso rus in the ordtr of 10-30C~ o of the veliicity of-
- 11. Lwa rssso m rus Ewrias S? Af f BUrt t>tNo- Oth b lB CT10N or LIGu rstwo Ft4%HR5 T O T R A Ns-bght. This v.tlue is attained as the current Psar III. J. H. traccarath. J. G. A n d e r son. us<sios trwss R. f t. Gmd e. .4 /EE Terrasotwar.
(IEE TrmaAss. pt Ill .P tser .4;;e !.d e l .a .M , :'i t *, ;p n.-10 becomes ci+ndant and is lirnited by the Smo vet it. W D 2 pp. 641 . p. iential of the strae divided by it s surge ~ imp dance. The voltage drop caused by a 4 y ,, 3,m.acu to m r Lu rxm Pr -
*~
the channel ferm.ition current being drawn rouxes or Tauswimows ! ivat C. F. Wagner. through the = urge impedance of the stroke ML. mL 75 Dec 19M PP- 1233 -M.
! pers the velvelty with which the Snal u 7,,pur .1unt, ,, ticu rm Lc,.m S e o s. B. F. J. Schon'act th e r.im t e .
~o rus 004034 L)iscussioft c a annels 1ppr..ach each other and conse. ,t . m i m pp 25-e.- - w >f ' '
qu ntly lengthens the rise time of the
- l
. urn nt it the earth. Ain ^ 9 ac des IGee-14 s nos Voi.%s sn ea s ~r 'N K. Berge,r (.beiation tr ciens, uurich, e t t ee rlind ) .. It t a, of a
compor.ents of the stroke after m, Finto. J. H. Park, H. N. cones.
. . .t tre,u
- o. m' a d c1 R*5e J'd. ,..satwnal Bureau of standards.
- r. or extreme value that the vithors undcrti og the urst, the 'eader f rom the cloud. advances Ta3bington, D C.. vot 56. no 4. Apr. W56, pp.
i a relatwely .mg - h veh. city (3c: e that of to draw a consistent picture of the Ughtning I at 201-24. b gh t ). bince the same cendition, that phenomenon on the hads of all wath 1 15 sixrv.Cvets axo turetsz seassoven or F. able research work about '.ightn;ng and in gn e rise to the h!g>, . velocity of the dawn- Lansa Gar Sractsos. J. H. Ha genguth. A. I ward lead, ev.ists in the channel next to Rows, W. J. Degnan. AIEE haa ns pt- comparison with new treacurernents on the earth, _t t is re,.sonable to suppose that I t ! .,( P. wcr 4 rtcoew and Eyrc<=0. vot 71. Jan. breakdcwn of long ,whoratory gaps. .This the return channel or. components subse- I Nr-. PP 4d"- Job facd. itates the wor.K uf the pew er e ,gbeer N.racs Gnatm uts ax !.c u o Gusot s quent to the Srst likewise develcps at a 16. Caisuts. J.15. Irigha m. J. M. Meek. rm<<d. po is unable to concern himself with t'te I more rapid rate than the return channel =45. r5 7mcal sonec y of 1.nndon. London. Eu s;4nd. e:ectrical details but who is irterceed in f for the drst component. In consequence, voi 63. Pt.1 Sept 195E having a clear. idea of the areas vi greatest l as observed by Berger, the rise time of the disturbance in his high voltage tr ans-17 Tucws=37va us -Tus Etacrase Frsto current at the e.trth of components subse* Verm M u so ruouEar Lmur m o Dtv m:ssion system. It also allows the spe. i l quent to the 'arst 15 much smaller than that en 4 as t s, 11. N orin d e r. report of the cialist to have an over.all view of the 'ight- I second muting of t he Joint Gn.russa of ruho. ".m g pr'icMs so that he may dig n er real of '.he nrst comminent. " M e tevr edo cy, Aug. I G- t $. '951. or apparent contradictins in the pict tre.
- d. A met hod ha.s been presented. for ,.
18 Tus .= A r t a s U,D \. W AllON or A f WO The report is based on measured or b-mc!uding the effects of the preh.senarge enanics Croso of Licu rmc D scu ases. H. sersed f.gures and dc.es not lme itwlf in currents of the insuLtor string into the Nonnoer, W. stdregen. A * *i, 'or it :rva an s.
***8 ** Frd- h k h"I m - 0* eden, voi. unknown microphysical proce".es computation of the lightning performance 33 A. no 16. M 60. pp.1-44 in the compamon paper the .iut hors of transminion '.ines. In this connection haVe e Xamined present knfurmation on it will be observed that a knowledge of the 19. Duit s wtu now or tus T A% s fwONr Or C"e nt sis r uow Fieto M sas- bre lkdown of N:g gap % aud adled w me s olt 'ime daracteristic, alone, of a gap bu ta sms.so wr es. S rnus C. F.s ,A a gner. 4/EE . ra smi =s. new in v e s tiga tic.ns. T1.e p r e se n t a n .n re-may be quite inadeg.nte to aptJy a pro- pt . III : Tace art.s mo a=1 we. M 79. suits in a tw'..part devel ptr.en' '
r Wer tective gap or to premct the f erturmance oct wco, pp ;st-w ordes which is composed of a metantly of a string of Un,ubtors with its grading 20 Rhevatoo brana rtssa ron tis In S rrries growin g corona disc harge with a '.a r ge rin g. The channel currents Inust also be or Fiat.o Cu suca Den rn Licarsiso D:sca ncs 3fet<*sgy and diameter of 10 to "O meters and a sery M. K oba y a s hi, N. Kita ga s s. k m.,w n. Adjacent objects such as the tower curhuus. Tuy , Japan, mL VI t, no 1, 1957. thin channel of 2 to 3 mrn of o that up smis structure wdl atiect these currents. pp. 100 -46. m verv rapid steps of 10 to W mtters
- 01. 1.n.urmc. D. J. Maian. E ruras c ar, Lua. 79 h. m e vhitv an
;un. Ec Ond, vol XVi!!. un . 0- r$ s-69.
gradie nts, tnd currents if trtincu' - i p i K A. References 22.
'" "' mth th.+e of la.rs and N- e e in Licn t:rMumo xo REawirwrstracur Dirtso
- m. b l .n K .454 ON b ei g ef.
4 0$.*(r8 3. bhUwi=E that theIe b a mIbrum N br lh, am Morst EAN SA1.VAIOPs K. or agreement. Tns Licu rNfNO S rsva r. C. F. Tagner, A. R. 5chwcuero.ct en E'e k tr os echa nc he n veretas. Zur-1. 41EE Tra mams. pt . tif 4Peuer ich. Switzerland, vol.16, nos 5. 9. M5 The transition front c'irr ena to irc dis-H P.e ma n. 4.; uw and Erdsmn, ml. 77. June 19M. pp. ,q ; y, , g o,, 7, . m ,yo, L gg, charge of the leader is c f "im'ul inte rmt. l
'"
- C. F. Win gner, G. D. McCa nn. AIEE Tro m aawes.
The drst cu rrent pip re tes the ,se ( 2 be Nr ru "r on Nsc i s nc s to Arc, vol. 61.142, pp. 9 trgo- Arp of rather em mt J 2 S-91 4 d .. .u t C. 7. L ner, C, M. *.a ne, C. M. t ear. & J . p p. o
.sv i m. ,A. , ten < cts ,
y r "a 6 a n. On l u r et.s e Discunnos t.tro.m k a es o rta k qf h.
" t ' i ' '
~
III:.t. VNim es f urt. A now ss R u otn ro inn m# ,ue d me has rt y 'ad a - i a WNY "OCW d ' d ' ' 59 T H Ar g Op rH g Ut T A6 s T O a, A. A. $ n0p 4 0, V. .
- 3. 'Iscarstru or Se A s trO W M or LADQn Arbire A. 5. Tor on a n. Es pors no. 411, C Ars. C. F. Ta gner, A. R. Huema n. 61.see P. I.anonov, in the center of tbs charge. Thn channe, pp 60 6-22 of this usue).
CICRE, Paris. France W54.
- y Wager, Hi.\ man--The !.ightning Strokc-11 Oc runn WH I C6 t
- t
1 t i
. a dis, pp 104 -tou. Isn. mi. 19, pp 191- h : a conductor seem to be sig,iic int althou g'a st. im Wm.tdi .tely and roume s later in the " " we m not able to back this up by cal-
"" "*P.
A solu tmn should be found by the N h. 'pp' s .IN. culation. 1
- N"- N ". !"j"n '?.12b . '3, .'N7T PL i 'i'- In the case of a high tower and with did , r .;- s pe . Jiet; to explain the reason Cerunv. vol. 9. pp IM M. sh.; ld condur .r3 orie n tel bev n! the 1
.e. -
t,. ur .,. tra vi'i in of ?u.g h n.e u
, r m d.c ta 19 low Celd of the 1 N s O ' 7.7.
- 4 m. "M. ,c .M" * "sr'
- u. t . i c e' wire, w hat 1 e4 the Ie ide r. cor. > e:ne i c.
.Le tE: w
>r bc A he, *he
- a;s c e divhar e % hen m . y . c c ., . , n:1
'M
.rb , !;" rc ot the ries the au8 r> 3, ,,g g,, pr M s ! wi M a An u the m er c- 4 73, y , g, , ,
of the earth an l rear,nar ly J t: to the spe u of the conceneratina of energy an t a W rm ate Parm er L Men, twb. the hem ng of the tine channel or of a 1% # An.urevi: New %r 2 N v. line ' It wes the earth tnd it & sees
.nr. w hich accorE:g t , Druce. M' M 'o the line condue ,rs as we:1 as tha hiehl s r md. wire and t< er perh ips a< t hu:L Cer na is i; gr s :tJ e 1 am pt n. In the cianical frura a;; . h r.a . and
, en ma t-
< > i P' h e M Tor , w h > ma ta or stre t the 'G, --HC un E. BG C P: Pa ' the l,ur e fr s .
in ; c' i ' * , o ' . b .-
.4 .,f Th upt pn u , a at a e cor,nt em e: ;e or tn > Gri t ecr this century at the Lri m
. rf go shieways Photvap s are enant ver O of Dr%!cn.especia;1y on Lici:&&, ictum str Le. It is acumc ! that the showing alrnost rii;ht and bnh m 1 ti g m 3 3r C'- n:la.b g": on glass plates it several recent papes on this <ub.ect have
.va s fo und that the tran3iti..n from shw lis. been motivated by the unnpected :!a sh- str 4e The str Le has t > dec de ancher overs th tt have wcurrel on hig% voltage to hit the earth, the tower. the snieid wire, chir, a s divharge) to are din harge iz g. -11 took ; ' ne, w hen a eer ines .:n high ! mer, with sin.;h - -!d wire or the !inc < sadue ur. Ths du on n:27 l t : t r.;e 5: pt cd wid i 1 At the eu! of the pr '
por re t xe be the e * ( r e ! - tn . . -itinu mi > sme is :n tde to th tt tran > ni, n ; ne pr i fe :in us fa:* r On 1 > ! Rv e ur:n i v ' in a sin; e nrm t strr e r.
' rs u t: be M to ' d IN :r;t chirge r: unts to appr ui. In +he tre.s nizi .n :ine pr J h m it hu ony of the e 1 to us that the prof. d i ay . f Ere kv be .x rt 5 p+ *t d a h' , i'i an mai 1 e.e tr etatic unit. -3 Ate' rd a g ti) se -
added attra :ti ,n for the strie. What i Topier. the :ightmr.g ducharge is a g;iding hits on the tow e r m nduc+ rs, so-ca:Te l I dmharge like Lichtreeg Sgures. but is in shielding failures have not been emphasi7ed we tre trying to say is that u the tower l surficiently. They x e not really shielding height increases, the shielding is less efec-sp.m . inst. :d of on a surface of insulator. tive and direct hits are promateti b3 this ) Another proNem which has not been failures because the shie! ding is probably is why not pod und r the circu m t,.nce s en. and by the Stas cause 1 by the line putentbl. < s- F ! b/ the disch try physic. 3 l our intenti< n
- i arc; n un>t th . -W ormei channel suppien ent cauntere L It is
" pr< ,
<t
! str. ke 'h- ry r: i nii ' sin l
it, r 4 imo, ii tiely af ter format; n of a la cv W ng b e he ' C , e r l'.t .rt that 2:1 of the une .pectc d hi + > are cert on :ench. The authors say that the theory, wi:h the '.a r ge car;na en.e!'pe. en m el cannut expand beyond the space and the Gnum the<.ry, with a bulg- of the rnu:t of direct hits 'o the :ine conduc Accor iirg to T6p;er this is not charga w hic h pro gre sse s by succe wis e turs, but it is our optnion that a high charga true . He found that EdueWy ric:res bursts, the chanas i f str .ke f ntact with percentag- of thern tre. ew m far out of space charges and he sta , preuy that this is the ea* with h.;m m ., Accordin g to &hout in 1, stop. , f pa z- u canned by the :le!d.cre ating action I ,l of th- :"etric charges of the prncee ting I ,,i l' '. j*/ i ehannel w hich corresponds to the theories i ' l of Me i ind Lwb on treakdown of sphere h^ ', i
;f n+'
I hase .e rised this e:!et t on the buis of h V f
] i F i .: 1 nl 6 of the aut%ri paper. Indeed C -\ {
the .r , nt >f a Mrneter space chargeq ., 3r - , j ,.
.i th le a s tie s nu .t ec.i in the paper ,? , ;-V ' ;:l." j f lajI l ,, j ~
;,'; l re uit in i value of ab..ut 6 kv cm e xist.
mg in the coroni space. 2nd theref.,re can d i ' 'd l '
* ([ p" Ol]
Ut " Q "_, The velocity of e:ectrons are g 7 I can ' :t.
%n, e re d creating a pluma, which is f g g A i
i
;ust the v Ng.ited channel The only a I ! i #\
j I ner ury condition tu achieve this edect I '\
"'W- m' ' ' ' ' '
'N-is that the supply of electr ms frem behin !
the tip of the channel is 'ess rapid than the g ,3 ,, ads ance of electrons before the tap. [,g' I f.sund the co npart .o c,i r.he snape nf
' f fg[ p. l eurrent fr'nts of art:6cial sparks with the S.#jg" .g f34fy ..Y W h. ?
fruat ..f lightumg :urrents of great interest. *L.,, 4 g igh, .;C i In&c ! t h.. concordance is astonishing y pod , Furthe rmore, the explication of 4 [j;@[..'
;s '"* <',-
t .. O " jr;.
.<-N
, /p ; w
. .y
?.
}
hj . ,/'
#/e l
steeper current fronts in mb~cquent strokes. N a s oh e r'. e 1 on ,.!. .u ~.t aan N h atore, ,ay the $ %,/, s, 9 A
-I # e 4
/t l-higner spe ed af the sun pent leaders is @4
;\Al.
s of. much mterest 4 g;!@ " The ,ath< >rs stated that they at+empted 3c to preg t a picture of the very complicated a phen n.st rt of the '.ig h t nin g st rde . I /*9 dM d h' think tv did this m a very clear and M j evocise .nanner, not as physicists but as 3 I g gs gs e n gine e r: who are accustorned to making n .c. ail the necessary observations and measure. a .cs *n-ments It is to be hoped that the yet ,, n. m g unanswer d basic physical proNems will T T ako be -Ive i by dise: ge phnicists. f{ t l f REFEREMES
,w eipo s.
fig. O. $uCCestive positions of feeder head 3pproac'ing n a to.er, A-D, and concept of pre-
- 1. M. Topter. A n aWei he P%ish. strike dischtge process of leader head, E-G j Ge r m m y. v oi 65,pp $7ks74. vot % pp. MO-
% Ier, 637 OCTotWR 19d1 Q Hilman--The Lightnin Stroke-U
s.. .
- v. '
g) 4 which is another way of saying that he j , ; , Fig. 14. (A) Geometric description of an elemental did not have sufTicient resolvm g pon e r [ d ' A [,n @[ e to deternune whether they tras eled ut
'~ N hoop of charge used in deriving 1
O or dou n
+
f , eq ua t. ion 2, and (B) the It is not too difh. cult to explain Schon e' ,' w 1 g !
/ general appearance of the land's experinne.; tally obsers ed "broademns i
! f ,,
space charge at the earthward 4 ,y of the upuurd part of their < the pilot luier { 4 -
^ extremity of a lightning leader t ra c ks." When the step oi curs, the newl-f
* ) .y . , _
3 /N ~ ~ ~ - k',j,, head.'toNoici The z.: plane formed channel demands charge, that the Ow i f t urrent. That this dem md i t ,,g , j / ,,, is about 300 rneten above t,
) the earth's surface viwnt and ac< m;um..d by a largt inert
'k ment of current is evidenced by the ph-tt
}
--s- graphic brightness of the stcp. and the re 6
$w.s i . jQ __ 2' i!!umination of the plot !e eler track. b J
l e n ginee rin g lan gua ge, a sh. rt duratio I hii;Str.a.patude pulse of current i- dra r a j f r .m the piIot lcader core and the pub
. trau b up t.. ward the cleud as i t ra s elin I wave. After a inhmn, the pd .t rade c.ae acts hke a transa.iv ,n .ne alt h di th ,u n in F;g 01 of refaence 31. , ,t L re inhin e i lo C, ind R denwr t s Rn ouc i S. B. Griscom Weo in ;h ouse E!cet ric C ,rp. rati .n. East pimburgh, Pa.): The repr. .doced. and <tiges (F. and ' G , of '.dgher pe r unit ', n3;th than for a w Jh
'"nductur su there is cond aNe w rg a u thors have iniq;ra t ed nany items of the authors' Fig. 7 experimental data and t heuretical argu. loss. causmg re.iUuminatisn of tht pLem j
ments into a complete theury regarding One of the dissimilarities of viewpoints channel in air, a spreading in time an 1 t the phenomena of natural lightning. In between the two figures is that the authors space of the current pulse, and a reductio the section of their paper entitled "Com. have asumed a vo;unietric distribution of in current an*plitude, as the pulse travei panson with Other The. ries, they pomt space charge v a ry in g i:n c r~ely with dis- up the leader to the cloud. in its quest f. out that their theory agrees with s-me, tance fr"m the 'e.uler tip such as to r rm! ace charge replenishment. This 'it s in per I ut di<,o;r es with other, news exprewed a radial gradient of 7,0u) volts per em. fcctly with Schi nland's +tatt ment.2 ' E ,c m t h e se the nes. The prestnke theoryi The pre <tnke paper used 3 n00 vults ;4er bright step appears as a ter anau. n of u ritten by this discusser was one examined em. This is imphcit from the term -3 X fainter streamer extending the a imle wa i by t he m. While agreemg with some 106 R in equato.n 1 of reference 31, w here from the starting p. int of the disch a rge ; elements of the prestnke theory, the R is in meters. Also, in desenbing their such (fainter) stre arne rs inc reein g i authors dirTer regarding tertain crue:al steppmg mecha nirm they wn9 der the bnghtness and decreasmg in wuith m the p . int T'is = .t hea!:h y ind deciraNe plasma channel to grow fr .m the leader carthward ends are appru hed.' usepon attuu ie. for it is only by the expresien of core to the b"undanes of the ' pace charge, tnat Schmnland was in error m .e views and tounterviews that the secrets according to their Figs 0,3, and 4, whereas that the startm; point is at the (!mi cmt ' It is suggetted that the auth"rs' concer J
. of bghtnmg eventually rna y be wrested the prest nke theory calls for a cas itation
! from nature. I std1 firmly Snhl to the of charge with a plasma channel rinally of the leader and its tip, or he.d, the f vali ht y of the p re st rike theory and its making a junction with the leader core, Fig. 2, is n mre nearly what w ould I
; imp'ications on the stroke mec h a r. ism, w he r e u;)on a moderately heavy current e x pect ed from a continuin g corona li there fore, most of the comments br. ught of 12M amperes for the average case charge, except that with the latter, 1 out in this discussion refer to the ddTerences h vs down that channel, formmg the stdl air, further (harges a rth .h isit i in concepts bright step and re illuminating the pdot s aryi, g as 1 r would i ut-nd hn . nd tt j To diu trate c some of the area, vi ., g re e . Icader cure upward of the junction point. a u t hors' a r bu rary b. .u nd a r y a h re t!
, ment and disagreement, Fig.14 of reference It is cunous that the authors thould gradie n t was a%umed to be '.' on ol 31 of the paper is reproduced here as Fig 13 shtnv the same prestnke cavitation process per cm. It w ould see m t hat m t he s i. .le!
i
, for companson with the authors' Fig 7 in their Fig. 7 that the discusser ch.>wed stepping of the natural lightmng erh j
.\ gree me nt wi!! he noted in that both in Fig 13,but otherwise desenhe the steppmg the phenomenon would be a e r..na burs !
j figures show bulbous ends in the space pr.(ess as a plasma channel gr..wth in the or impula corona, where the thargedil
' reverse direction, that is, from l A) to i C) pl wement is rapid, producmg loud t rxkir i charge volume surrounding the ! nver end ounds w hen a wnductor is sud b dy ~s e in their Fig. 2. This they just:fy by ,
of tne h mier core and the authors < tate ei+ewhere that thev attr hute ner three w. tin g knonLnd? " N *im t e e vi le nce s olt ed The unarge is mmtly nc; ' that the streamers travel d, wnw ard is. trated in a -he!!,tutste iif w hh Me t t(tr
$ tunes as much charge per umt :ength for the tip as for the 'eader proper. In their however, afforded by the brmdening of gradie nt as era ges 30,000 s olt s per em j the upward part of their tracks / However, siigntly less.
y pre uous paper of the same titk, the &honland was refernng to the pd.6t leader, While the authers bas e now recognia f niathe ma tical analvsts and conclusions and not its extension, as evidented ny the that the charge on a h oler ca n not t f were entirely confined to uniforndv dis- fact that the foregomg quotation was regarded as ca stant per unit of lengt tnbuted charges along geornetncal lines r they st dl u se line charge < to
'* * * .n e n titled "Jeb rity of . .n +u Thus, ther have adopted one "f the com- '""" I"'S t re ame rs, and also f rom pre-note n nals. TS is 'h. .w n b 5:
Leader * ' : m-ponents of the pre 3tri.ne theory. Further ceding and foil. wmg c.-ntext in the <ame 4, w hich apphes to h.ne (h rges Eht s pomts br. .u ght out ongmally in the pre, . paper. The following <tatenient :s made prortdure is tn use z .ne- .f r a ge wr x . strive paper, and used m the a u thors in the discusser's pre <trike paper: . .r h is , the lineal density of charge mere aing s
, p. per, are: analysis teads to the conc!usion that the the ea rth ward tip of the kader is a 2
1 The hulhous wlumes of space charge pilot leader trace and the bright step start proached The ma sitnu m rath. .i time mnult aneou sly md at the same point in charge density used by t ',e m 13 17 % ~ at the owernn4t end of the :cader are n. pulh n g ch ir ges of p p' wi t e gn fr'an - pace, the rmer t ra v e!!n g apnard, the e 3 / ~. t o 10 There can Se .o grrunded ..bjects p rc.iec t m g u pp ard Liter downw ardJ to the u e of z mal s ar. , - .n. tp It appears, therefore, that the authurs unit len gth a ben t he ad uw Sta l C. There . re ca p& rial spark t rac enes and I agree that the <teppmg process is the s olt a ge . utnbut u.ns t.. < ate.
- casitatmg charges out of the leader head self. triggered at the 4 .w en nost tip of the remote from the :c.ider. H ir s er, t
' pace ch rge 'the dix 6 e r 's stage E, ;cader md not at de mud. and t onse. !!ne c %rge :pt d.4 < not pve '
and the autnors tagm t and Di quently we are both in dkagreena nt with t .r r ect s on t. .ur f. .r t h A re J 'arp
& h"r.L nd on this yrticubr p. int A a be+ber the chara ~l ~ ou o ..we
! ar A
. t r u . nc s a .. ' e na h.re ..nt
, . ..ns'. r
.urt ie r i .t e a - m; p a at: .n nm w e .n. ~n the ur f ice . .i U i
a .i
..a p)
- m gg ,i. r na ne :S
'a nd is, 'the ups ,i tha e i ocrs the r v er: tion or .r. w - s r iu%er? % ge 6 and ..uthors tage E q. channel .uten ti. rs m en mu me eshdat dst mie .'t o n '. . e r 4 . 'in g .
i
- l 4. Cahnination finally takes place as no distortion due to lens motion t he authors' ci a ;,ts i
I nM Wyner, fli!< nnn -The Li;;ianing Suoke-ll n c 7, w 196 j l 90004 m
In determining the contour of a leader tensities of the order of 100 kv per meter T he followm g treat men t for charges for nearby strokes. Fig. 15 is a typical having rotat.on d e mmetry with respect head w he rein the major portion of the to an ad< is tnore rigorous, and permits the charges is regarded as bemg on the bounding record. Attention is called to the fact suriate, the nrst apprmimations are directd that mrt recchin g type triodn or pentodes
- c. .u t mr . 'f a s . .iu r of charc: to be de. w iu n.st then 5 n :t, input t.m tr i obtei g 2n eqaqise,ti t! <urhee ' pt i ter i to vi w < u ic ;. d." .
r a:.!r TH- 5. m t they by tri ! md < - ~ r. n si the invc :n;<' r Tbi- me.'- n wn% ' ' o!. i up p. :a i wil
- a ' . .r t E ar . .t n m t u. ', cor.d ent v >nt and h. - P ,
in
- i a st' < is ut,- a4 diu Mat.-i chargr dem tie 3 A of equ ci m ;.' t . -uch w iM fa H ei ter n al th sa ci in Iis 1 F B . !-t ll ini igg i U A .he .t cucular mpu' v. t a .;e in a mu:tt . a a er, h . .. .p . .i e m. r 'ocated m the :th lamin t that the potertia tt p. e r t 2 ' . l's. ?c i for ! @ m:;
m o, and pa d! . I to the xe plane Let A be etc. on FM 14:Bt are e-entiity eqa d w hkh i, let t w r,
. etc . be pom t s hca-I the!nt w , u - io he s an my uturate th r . i r c: de 4t> per unit <i Let P J. Pi, l'
' .i5 Cis '
m u u! - by s e r, t ..! to the rurf eud u wr ' ~. t n ime rl
'e:. s h 4 t h. 4 . ,) .h per :.:"
l nn r twc
, utm as 'e :U.cd <n Fig di o ila .!! eh m -,i , me c A. m i"
- i n:- % .
;i \ T hem. he e .n t r d n t. , tht a bm . ,,tn l D- 1+-2='
T he
; _ r.ti d m wits d ii or a p.. tnt l' m the 02 term b w .eut tilter" a unic: i .:t 'i me, pl.. < hie to an indnitemnd kngth d l'- dl = d i' and 6 de -c wt i. ai m.u:.1 N tp m .' d of corge at P is s he a nie e scur,i. .n utj ,' u ttan su then
. ,n Fig 5 ree v ' lik# to be lae to q ;. 9.w. Ja '"
dl . :'3 the cha n n .; nd sul 4 ' r-c .e- i d: - - 20' J r al = - ,
, pu' _
,5 vaca i tube ,na ' n. i ,i t r of oh y with a < r inh a of
. wh.m 3 he s t , orc ~' d ' i9 !
. 1
- c. .o J inter-t a g s 4.c ; ., c a ; . r e d ir p .r ar ' c- uti - , et
. i 4 po at t: ! ., . u? r at: at th_
gm .e t. r + tant i , I un 4 a,. 13. ut due to the e tire
- Op a, ,i is .,btane i a4 corre pWing unprit pu n t .
tun the mte tral ct equation trum 9 " O If the gra iient d 1. d6 .ed, given by equation the value of ini. rmation conce rning i s ctric, to de r' This is - 3 is not the desired value, th en t he hoop 6 elds due to di3 tant thunderstorms as charges mu st be altered acentdingly. It recorded by vacuum tu be or tra ne tur shnubt be un de n too.1 that the leader circuit r y One need only cont em pla te WDd< h?. 4 h 6. n .c it. i Fg 1OB and 'he thit thund-f orm, m are t aff un ind et be gn e, d ue da a < . a t e- .nur ain
- cnM . "A th i s l _ p ,m ! .,
van ns im y uv Fi " a t vJ
..I azin atha con 3iderat. ct in 1co raa: atwns .f the it se s : tr leaders are mu:tt;& hr i cd. 1d the "h pot en t uis of all points Lme or pomt char;;cs can he used w hen D: ts large com- zip an.1 a c .tre rana:nh te ard way pared with 1, Fig 14 L A > In the idealized or para: el to the ree. .rd n g < t at .. a a ntf M .
va rt. iu d y polannd and re:!ec t ed . XI of
- k. f ir '\ M case, the lestred surface 6 eld inten3ity is pg oit s per em, but irre gu;a rities this re3u'ts m a con f.n. ram.n 4 puLci Dtooo and 2 vli we 'he nerig cidenttibe nu, h arrn in 4 at the rec ~rd i" .; s t a t io s -
catts ni wns a
'e, Any p. .m l' . . l' . ', e t c . id F n; ledu o a - as to tun 161 uato,n 21 is a teimplete ei!iptic intep.tl 14. B - mt. t not be ,it .c u h n t uth pomt , !ead cumat fiv ef ..n nut ir . rd mu t pr c. uh d N k < t, and may be eva:nated on any hm ,p of charr as tne e!!!p t o: he apprai+1 mth extn me cautu ut I mteral mrt:;uiati= >n is not sa:id for :ucn have 4 taine'l % .re han C rece.rd, of by the use of >tandar'! puMi;hed table 3. elec t r a neid chan g , due t o thunder-t rms The tot.il pc tentul at any pomt l', F. g ca s. , if it is de-trei that thew he a ; na!! '
l for any surf ice .1r volume < i t har.;e f ail ..i po rt di.J f r. aa 1. to l's to ?- etc . at thre n e m.re ::ent 4 it: m m te or 14 A
" n c 21 about the y nts may be then the vectors J1 6 'c . c in be mdivl 'es m s t r i. .;4 t e ut --u - t 'ine. % ced at
,,n:
mined by sun. ming up t he sontrd u- from the normat For em:p if a fall 4 mde mte ai, The thrc 6 M .4 v. ' mt d .. 6,' n i dt, per tom, of as many hoops of chirge as of l ion siat; per em is des ire I t:< ng the me a,o rem e n t wa- ab.nt charr ensO pe in ! T 0" w ol t .+ pc cm meter UM the up; e - hi nt th i a W.4 0 nec ,ary to dedne the po3t u! ite'l spae [t a n e.um! ' hat hinr g chirge w it h rn pect to its contour and f rom l' a to P 3 ', etc , thea dl can be volts per meter if the enarge is direc t ed downwarih about 4 .ae rn many t han t. tms. ret .r f w ere 4.und cn.oge dnt ribu t o rn at oni c ne . f t h three rw r ' m; W itmns m iet. m the ,urf ace. then only one hoop from the n'.rm d with cati. fat , r :. Cuncy pe r tu:ana is required If the charge s s. !u- It would seem that wnen the 4 der ore in other w1.rd lr th s eaum e nt, the . me ne rily distnbute !. then 3es et al h~pi r has penett del to C of the tuth .r v Fig 1, masimu n Wstinre from a thundere! .ud the elettne deid intenotv at it + tip .u.uM w inch u nuM pr. .due a re- ,rf was ia the pe- : t mma : : ty be u>ed, with varuble hwp enar c temittes %, b. etc. For a tightnmg be e nortr eu, enrrparei n what it was order of 4 n. des Thi. 4ee n- N be a de-w hen locate I at .1 In the prearde the r'. D pc q r' leadc the adect of ima ge chargm in con-trd utmg to the potemial of ? must also a "vtrtual an ale" ea vi t a t e > carge from - ii D D f3 1 ,- M9 r,9 [.
\
.(, p ,
.;(
8 be t on.,ide red the leader head and this .:s a non. dent J jg ; ]; h , .; i
,,,,,, O proces, untd it bore 3 its way to the cader gym a This met hi d of obtainm g pote n tials due to , pace charges is not nearly as time- core Whde the author. ,how thn sime h bv.7d;; 3. g , E,;,,,JNJdT;M ca vita tinn procet, in Pg - where the "l.5Ny s -gfg"~g i on summg is : met be antic. pated T rul-anode a a real one, they lo m.t do u durm g y- ,- .- n;- p'"
ar..I -rror protedar , are necessar y , as ' with the line charr i ethod u-e.1 by the the st e p pi" ,; pr.ree and tha a ppe .r4 _- _' '
. t author, but c4 ns erge:a e n r ed inconsi*:n! .
1 They hase endeo ore 1 to a -e the re. . - s .' t e - T he aut h. ir s arnve at 4 h io* volta as the cont nbu t 0 *n of vnit a ge at pomt 1, cord.n gs of etrtam un e ,o g at. rs of mec trie . * ?* L-, k12 fM J 1[' #/{[ , ,. 4 their Fig t, due to the hemnpheneal deM ntensit y to su p po rt t he.r e r yon of . (- . .y C the '.ightmn g lischarr mechannm I ' tad . i','33* % p sz , boal if charr below zone 1, whereaa the a
. L. I x. / %, s .,
1
- m. ire ng .rou s pre redure even here 3how s eumined simdar data and enelu@d that 6, m a contnbution .si i % 1 08 s olt s This whde useful for certain purpo-cs type of record was unsu:ted to determmmg this n.f' M .4 ~ 5.y[M"; A:j7pf Vg 'Dl%A 1 ?
di&rence does not tell the complete story, the mechanum of the hghtmng ,trAe D -
- 0. .. ee ; Ud 5d, 4 ber.: u se it must be borne in mind that die proper contour for the wrona ens elope in their section on electne bent mea 3ure- I,b,~
gg p > 4y:g. yr.g"(h. g 4 ;;,..,@; a dt not be ontame.1 t.niew the computation p r oc ed ar e ; are correct When the more ments nest to the earth, the autn. irs peak i eieetronic deuce > has mg "high a m pii . yng-p[Q';y(,p*:ji,4 g.% g /Q. N ig n gorou s treatment is used. it wdl be f4 und Scunon and sweep res.dutton" :imes of M Tr-hqfM'g',Al' M'a W h 3'M I-t hat the form of the corona envelope shown !O psec. Hign ampiidc ation ap pears to - bv the authori Fig L does not adequ itely be the m s er ve of the requirement s int Klydonograph-type record of durmg Fig. 15. res eal the lisprop..rtionahty betwetn the neid mt en sit y measur-ments electric Iteld Change of about 100.000 volts charge m the buih..us cader head, md
- he t hu nder-t orm s Some of the discu eer +
in ve sni;a tion s ' >nowed eiretrie de!d m. per meter during thunderstorm channei charge immediately abine Wa g rte v. H O W llW LigOtiq SYNE (( W OCToHER 1901 i 90004037
4 11 sirable attr:bute. Tha essential point here until the gradient at the boundary averages paper. In conjunction with nur associates j is that recording devices with sensitivities 30.000 volts per em, or slightly less. we are working on this su bject at the capable of showing 6 eld changes as low as present time. As a result of work na the a/ n.
. Because of a random protruding volu.me -
1 milh.v.;1t per meter may produce records, stroke mechanism we have not only rnad'- particularly at rught, of atmosp! eries as
. of space charge having a beal surrace -
1 s gradient in excess of 30,000 volts per cm, tied our ideas concerning the current at f tnuch as 1,000 mdes away, whereas devices plasma channels cat into the main vclume, the earth and the movement of the charR05 hav.ing a range of say 6 to 100 kv per meter above the tower, but feel that a ogni5 cant will most certamly only record local events. ca vit atin g adih. .tional charge into the protrudm.g enarp voame.
, , number of the unexplam.td b.re !b h m ers The authors present Fig. 9 as though rn the 315m, v lines rnust be due to hielding they regard it as the equivalent of a ' map. 3. The pluma channels f.nally eat their f tilures And so ne arree with \f r Beck t
shot" of that part of the e!ectric discharge way to the tip of the luder wre. where- in emphasing shield m g f uiures as a
; occurring during the interval between the u pun, since the energy a vaik ble to the p miUe ..,urce of trouble. , , applicatwn of voltag and the time of the pb ma channels is now relatis tir un nited. We opee whdeheartedly wit h Str.
I voltage chup by a parailtiing gap Actually, a heavy current La s d su n the t.h innel G riscom in the de-irabihty of the free j such phoing aphs are composites cf the e xt en+iun, illuminating it brightly, and expre eion of vie ws and cou nte rvie ws.
) happen'ngs while the voltage is rhing, plus pr.gecting anoths s t !ume of wx e (.harge. We nun admit that his further i mmem a j the events f ollowin g the chop. Sierrdt at a .oner e:evation rhen a perdid of hase not muditied our ow n upmions re .
tnd mn i!!ppeP eonclusively d . i "ated reLiis e quiew e nce bMoas untd an.' her prdmg the ; hyscal procee es involved in I l that tiectriinic charges propagated out from pr quher mee causes the c uence to r peat the ly t nm g stroke. M r. G rium has j an electr..de into a gas will rush bck to itw!f listed fuer , eints and has prm !"d t he i the eh :tr .de when voltage is rapidly re. u it h n additienal . "e in w hich he m , *im
- y. , . , ,
nu <n m' T ed f rom it. They term the reuitant Lu N derptype recordin gs as f "# *3 "- "', PT E #j""5 * "' '",' d, O that t Fe<e characte? tics of the diu m a ge "bac k. long way waard an under<tanding of th."is ure on ginal with the pre-t rike theury ~ 4 C gur es." Park and Cones' Fig. IP show and that thev nere ad ,pted from his I identical results occurring three dimension. m st fascinating and important subject. previous paper' on the subject. We have' a!!y during chops. The upward projection I w uld like to repeat the appeal made .in stated that it is dif5 cult to determine the ( reference 1 for a standard terminology for of charge from the negatively charged wurce of component aspects of this com-sphere is followed by the development of use in de,cnbing the varinus phenomena picated phenomenon and in rer!)ing to downnard!v directed plasma channels from of e ic Meda n ,a g oes. It may
- Mr. Griscom's c"m ment s fuliv de -h mid
, both the Irmde plate and from iucal con- that nany of the disapee.nents bet nen am. to re.iew all the l iter $ture hearin g centrations of charge w hich had been' e due to mumterpretawn= of en M.cse p<ints. But irt c mp iring our O'""5 i/n c! min g. I unuld ok. to ac-i dtposited m the intervening air space as terms ow n pub'icati> >ns wit h the prestrike paper f the vdtne ws rising The direction of knc wledge the e tance of D. L Grucom n e ha e the f-ilowing comments.
tra vel for each of these phanomena is Ir his helpful sug;;esti .ns, a r.d for his it is not <orrect that in the previous un ques tion a b!e i, view of the mass of devel, pr ent of eT!1 tion . l. p per / reference 1 of the current paperi k evidmca he many invro w *< % t the t hat the s + ic vac < -t! Hy + fmed t ra vel d:rectiva of gueous electne dis- g u ggseg3 to unib h & ribured t.harges abng charges is revealed bv branches w hich gmmet nc anes. In Fig. , m the previous ;
' *, # "" " 33 " * * "
diverge in the directic n of travel. The , aper, the estimated vertical charge dis- 1 fort going is wit h respect to the small 1 m merme 1 of he mer- tribution in a te.Mrr is shmen. It indudes I electrode as r.egative, as is the case with 3 See r. fmnec : of the paper not only the end edect but also the edert
, most lightnirg Icaders. The pr.. cess is 4 scurue cuws H mn FLEc: pic F;rt r, S. B. of the vdta ;e dr >p in the central channel slightly di& rant when the small electr yle G osc oin. C.m a. Leu. New Wr k . N. L ferdmg t he t harge. Laterally thr.,u gh the is po<itive It is of further interest to note ^ P' 1R M1- crms *ctien of the !eader it was recognized l that A.A' and B B' in Fig. IS of the <ame 5 Tne uw rmycw h rmor.rm oe that the charge has a volume distribution. l reference, show voltage chap;)ing to result Wu n oc Fi a wu m Tmv U n;c u os In this previ m p1per it 'vas i samed that l
! in an ilmost identical, but reversed dis- Q[y1 ^5M fj[y"T*[$ ,
the boun6ry of the charge was such a
'; placement of charge as that which took mm *:c. nr k . N, v., wt w. o,c m3m -urface as to p*ess a gradient due to the ,
l place d urin g the application of volt a ge. 6 see re'erence 14 d the paper, internal charge of P,t OO volts per em. This near equality only exists for discharges This is the snme arumption used by
- u here there are few or no plasma channe!s SN tand an<a Crisci m, and probably d e v elo;,ed The former condition approxi- others We ecognized that an internal mates the char;;e and discharge of a locsless C. F. Wagner and A. R. Hi!eman : We gadient mu .t ewt althin -uch a vdume ,
ca pac:t or ; the evndition ms oh mg plasma would :ike to thank Pmimor Berger f. r and e-timated t his r$cet by umpiy sum. channels is more like a leaky capacitor his cemplimentary cements, par:icularly ir:g that the radius was half of the nlue I wherem a large part of the irutially stored with regard to his c 'ncurrence atth the so ec mputed But w hile it was rec 9;nized i energy is dissipated in Jou!can heat and is a uthors' engmeenng approach in their that the space charge was distributed mer
, therefore not t :.ot erable upon discharge. edurt to corre! ate the stroke mec ha rusm a volume, we never attached the same The authors misunder<tand one aspect with laboratory measurements on long gaps. importance to its precise determmation as of the prectrike theory in that they state He has rai*d two sery intere< ting points does Gridcom. While we hase modified "the charge volu me grv.vs. mint a h:in g and it is hoped that iurther work on the-e our own th nkmg from the a "u m pt ion all t he w hile a 'urbce radient of 30,000 subjects will ren eve much of the igwrance that there e ci-ts a fMd of 3 Um vits per volts per em. The paper dr es not connected with these phases of the problem. cm around the periphery to the a,urnpti"n make such a staternent. It is only when M r. Beck has introduced an important of a volu me distnbution that v 4 ries n-the surface gadient drops to this value, point that norries the transmiwian engi. versely as the distance from the conductmg or somen hat lower, that growth ceases or neer While this paper is not concerned core, we do not believe that either one can beco nes very slow Fig. $ of the pre- with the application of the theory of the d nen be the bour.d try precxly The strike pa pert plots charge prop gation stroke. cert am cha ract en stics of our actual phs ocal proce- es are ,.e %h y mo rites minst go dient ratio the aumber nsch e n tre .hrectly , Vable to the o mp.ex to he at: Ced W -xh /c of times the surface grad;en t esceeds proMem of the hicidkg i cor.duc .rs by . -u: pt b ms 30,000 vn!ts per cm t. During most of the m erhead greund aires. This
- ht ury indi. Wit h r eq. c t to Cr 5 mi r ~t . m t.
growth ai the Icader head, the urface cates, as i dlustrated by tquation 9. that reference m ed only be :r ie to h; 10 gradient greatly exceeds 30,000 wits per the decisive po nt in the pr< pess of the of reh rence 1 of the paper to :ndicate the cm. the amount de; ending upon the mter- le.oier to earth as to a htther t he g rund rec T ton of he pr. :act on of erwa play of equations I of the paper. wire, the conducwr, . r me rint < m the ;m rge imine-d by 'he d. s d To i nfy the cader o cprin g pr < ms e art h. a dl be t he nate -tr ien - ,m t . ' Her . in s in ' ' ers >i ' r vf he ,
~ t r ke, the in ' ng .ree e- 3 01 ' it be t s v! +f>re a '
u r i :n r N'r 15 , quece< are ep*wi nc-Moutmga iep. , ' .. ps F o ' r L O Tret ir m the < sen ,! a h be . ce . , e
; int is atta ned. This :s a lirect i ic. ,ne i .r ge in the d m nw std .it u .>my In bs Mr Grm i It h o ben k:v t.n or lea ! : core and earth prior to the i s er the uw of Aln a re, u s pu ,i the a h r; tin- that spae charge and upw ard But whde the #p. ice char;e distrd utM rec' >rd s of a mu.im a of 10 We was cha:n b develop from the groun I as the is under discunion, we midt men' 'n eben hm. thw limith at toe, not lcader appr. .a rhe) that we have been un a ble to bternune adtet the accuracy of the ree : -te ! instan- j G ris refers to "cardlarial -park the reawn for choo3ing the part cultr s in - ta m us uiuo of the :ieLi mtendit ! +r actn -' that cavitate out of the head . f my vo;t s per em h .r the pir.t We h; e in ur pwe- ri ar rd .brun:1 !
m.t .irge m i the sh -p. dmit .gg .1 . h ., , rm i -. h. m i r Dr. Br.4 in w hie ' a mmum na at vi thr eh2 :.ch T his !ar gu.:g"
- in tue previke the .r. It is r.idd ., - hi: de ' c tt. ' i vi he F i ma M1 withi t sornewhat vague and ne do not fu!!y under- the ma thematical 'e s pre ui ;n, and no the ,cale of the mehe 'pe Thi3 rec it i stan i the imprt of this phraseology. It men ti"n made in the test of 5% this prmide, a dirett companun botneen the is mr undentinJin;; of Gn a nn's theory particular numbs.r w a, denced mr;nitude of the 6ebt intensty produced thit the "cavit ati Furthermore, we would hke t a comment by the steps and that produced by the the ,> un !mg sur:yn" ate etprocess the >pwe begins ch.tr4* from m vome a pects of Sir Gris , is ci.mpute R change w hich cor.dr: 4 the S t .. ment ani prq .; e - int , the space charge a m vi th e c,ur of the rpa-re . i, that the for:ner .;re : J1in c - m with in 6 pt ,en' - d& nt fr em that eg g .pe at the imt.utt of mair o:n de. the !atter. %he th. e pi; - u n- !
pr. ! u+ m that wt be: ve that the y,!"; t u t. As sta te 1 by him in he for m . by Dr sta g m ; M T 4 ) extt . the chqn d dev >p r in of his dim m pm .;t s the ::igh ree, onx . .( '!.e p r ~ u c';un ' the tip of the conductmg core or the .eader eqation 22, the tir > t approsimations are the o3ci;h gr a m, of the new equip tent and ut the h,nal 3tep trom the tower top directed toward obtaining an equipotential show essentially the same results aithout _ alw. W ith this unde r3tandin g of the such defects n overshet, etc. surface. Now let it be assumed that the mea m g of "cavitational" pwe> >. uune leader head shown in Fig 6 of his prestrike Due to language di$culty the " low cut ofthe proc -c3 fexribed in rig. . or the paper is such 1 surface. Further kt it tilter" ments rd in Dr K: tag inis paper pa, :s a "ca vit u b .n.d' pr. , a >ut are N s M. v an ippr nima that thr rekr i N a hi C! Ta a . ute Tne mm
>u ;.c .e u. p iat at we >em2' esternal :urface padient i, ormai to the sh%t a h:ch caa bc < at Dr K t i;i .s a s na t m g lr m relatively go .d ondu ctin g g,fm wi .g w yg a l ya eld ree rd ,e otcauw!byrto . f r .m elec trod e, frem the leade- core for the and that the intmal surface radi-n t is a parlehm; pulse of voltan Depend- ,
downnard developmett and from the mast abo normal to the surface ar 1 equal to ing upon the input time emntant, the mer-tot the upaard moving channel in L M volts per em La if i mail esiin. shoot is more er len dependen' upon the refert u I we deenhed the procen of der is thought to be >itu it-d in sue a d u ra t i-t of the pule Dr Kita gan i, in , chir;;c..l4 ..u by the retui n >troke as a uy tM or.e of the dat -urf ico 'in in !c Jap.. n. u >cd a 3 nn .e. ?nd t ma. con- ' xt: a a:. c. , .a ; ot. y;,1 the other out @ . .i re n.L 2 unt in New '
>bd > E L";" 1 Mr. Gnwn objects to our u.e of a hne @g i .g g and Bri,..k u,ed a hee time con ta-t.
char;: to compute the potential when the Such us ershoot- are n. .t mig udeant. The l char,;e is m reality a votun.etne dictnbu- com .1nent of fieM. ar.uund this c) inder s equal to 4r times the surfxe charge system it self, an te nna +, a m phners, etc. ; t io n . Aith spherical distribution, the e . 3 @. . wu checked by simuLt:ng a umfe rm tield por e n t ui at any ppint is independent 01 abm e the antenna with 2 plate se veral the manner in w hich the charge is d,s-i distnb ited on the emeN a bie rt is ed stant over the entire surf e But such inches abm e the ant una and to ubich tnoutri witnin the radm > et the pomt a constant surface c.aar?e - i d -tnbuti~n upon a rosimatel 100.vo:t s u.tre was es acre u m,- . con m.,e ra t ion . us is true or 3 , lied- The !ineant y of the equipment
,he contour of h..g o of the prestnke paper s
m n:at e, v on g e>iin rica istn .utions. s t was exce!!c .t. An Sa dec: bel attenuator
.>e a n u rn. mM to the ty pe .,( sn irge is abo true fe the potential of the mid. g g ,
gg den ty dep,cted in t ig 13( A, or the same i point of the su: face of a cylinder of dnite Tb m length a long 2s the length is several times adjustable and the detting was determined the radius of the cy:inder The total charge Plf"' in F. ig 10 or his discussion. Sir Gn.scom by the dLtance of the stcrm. This equip. can be regir-led as distributed along the say s a dram 1 tic uvueWq ngure ob. ment has been in use ad was developed, nis. This asumption is valid to axial tamed by an instrument hxated on top to its present deicree of accuracy, af ter iengths down to several times the radius. of a bunding approximately wtt'un 2 mnes 5 years of que: toning sim:lar in nature to , i The range of validity of the use of this or a lightning strue. He states that th'i that mentioned bv $lr. Griscom. l eu remon is dependent upon the actu I record repre3ents an electric deid intenatY ik Grimm is'in error in his statement charge distnbution. For a charge dis- of the order of 100 kv per meter This is that thunderstorms, at least in New hienco .- tnbution that vanes inver3ely as the not surpnsing as it is known that coron2 wnere the3e observations were be:ng made, % radius it is valid for much shorter cylinders discha rges occur from objects clo3e to are targe-area adair. Thev were isolated
~
than for a charge concentrated on the strokes and for a corona discharge to occur and located by ob crvers in all directions surface. Considenng the present state of the know!edg- regardir g the charge dis. the 6d 1 intensit y mus exceed 10.%O voiti Fu rt hermor e, this recor I w as while the recordin g s were bein g made. [ per cm. Their c"mputati. .ns show that the prd a-tnbution we feel that eur u<e of this ob'ained on the rot f . f 111 baMin g a hich bMty nf overbpping of succesive strokes expression, in this m:mner, is justined in :tself enhances tne ma gmt ade of the is verv low. Perhaps the 5est e vide nc e. We wish to congratulate Nir Griscom on ob>er ed 6 eld 25 contracted with what accorEing to Dr. Brook, of the exi-tence of his development of the expreuon for the w ,utd 'aave occu rr e i on level gr-nund these pulses and the:r identi6 cation with a poteatial at a point due to a umform dis Unti; the method of applics a uf data sing!e ligntmng stroke comes from their ' tnbution of charge around a loop It is af this :ature is gnen one canm.t comment correlation of actual stepped :eader ph< to-a convenient expressian in a practical form. upon the value of such record) gra ph s with the stepped leader electnc l He then computes the contribution to Afr. Griscom questions the suitability of tield change. Such one to-one conelations the potential at point i in Fig.1 of the e!ectric field intenst y mea >urements for are very convincing 3 paper, and states that the more ngorous the determination of the mechani<m of the We appreciate that under certain condi- j procedur gives 3 0X IM volts instead of lightning stroke. Based upon the papers tions an arrested discharge cannct be re- ]. i M X 108 volts obtained by us. This of Brook and Kitagawa, and supported by garded as a 'sna pshut" of the electric . difference is less than 10% of the total correspondence and personal conversations discharge at the instant of chopping This l potential and w hile it might 2dect the shape of one of the authors with the>e gentlemen. is particularly true for applied volta ges of the space charge in the head of the we have great con 6dence in such measure- just below the entical value But for leader it would not significantly adect ments for studying the mecham, n of the applied voltagS just above enucal af ter the potential of the :cader core Any streke The followng comments concern- the channels have had an npportumty tendency to iccrease the radius would be int. the accuracy of these observatens are to de velop, the bn ght channels i high partially compen,ated for by the increase ba ted in part un these personal contacts lumino ity do represent the progress of 7 Ocwna 1961 \\'agner. Hikman- The Lidaning Sede-It 64l y
i i ., .. I I ' i the dncharge at that instant. Thh has indicated by a heavy line as the line be- Furthermore, the existence of such a pre. been venfied in reference "t We concur t w een a and c is indicated and if snost of the sinke current, a term from which the theory
, with the discuwer that to a large extent c harge transport con ists of the transfer or dens es its navne. chould constitute a natural the current that flows prior th the oesclop- the charge from the old head to the nen ingger for medlograph measurements of ment of channels is largely capacitative. head why in Fig 13 of his prestrike paper current ach as made by Ik rger Berger Hon es er, the fortnation of the channels thould the contact of the channels propaga. hac dis m ered no such currents but his trip.
is cuentially a thermal e ffect This is in:g upward from r with the tip of the chan- ping inas been - onewhat low so that it is dis u"ed in the p.t per. nel core at a to stttute a trigg. ring act Mt * 'uih!e t hat be m iy has e !"ed tr ea - are. We .ckrs.wiedge the error in interpreting If \f r Gn com is ei rrect in -h..aing i : Fig r eut <.i -uch curn nt if it did esict p..it
- he the Gnecum ne>trike theory with respect 13 F > that mut of
- he current d..a ' -m imuit is ertainly not so ion that -ueh a to the gruwth of the . pace charge s olume. a to c then m thW me there u r uk! be noth- pre trike turrent uuhl rave e.'nstituted We 4heuld have st ated that only after the irg cruical ab iut t oe not . int at u bich r un- a a d oral tripping desice for his -ello.
u.iume hau des e!uped toits maumum estent tact is m de at a paph T h nost estry record wou d uve I is the surface gradient equal to 3n,nn0 volts t. G risemn3 di vu ei<>n is unt erned .;p; mired as a t>e r fect record l { owe s er, per em. We ah,o appreciate the -uccinct ah -t entir@ with the mu h inism < ,f t he -ut h a as not the ex fh rger acMally 1%t
' ' m ements th it descriNi the ' eder < rep- !cader it i- i o r . .pirn. n t h at ti.e fen nee ch of the early erti en of t he " cord .
; ing process But these ( ..ttments seem in the current Fe earth ntniitted by *he a vitinti..n w hich we mar (tand is ! .ein g i ,t rangely at s anance u nh Fig 10 of h!r. presrike theory and that predicted by ne i 'ted :t d be Mue tna; to -bs rve Grheem s pre.mke pa per Fg 1:O F , t h e. .r> , ,i t h e aitY rs is of gn .ae im p. r . th . u!t c of 'aiure : . suruau:t ' o A ttr.
hows the 'ist stage of t he .!cs ci 4pment of a tance li ording to the pres ike tl wr3 viine the nat ure <.f t he hvitning f rmits. new step At t % .stant a ne, -pace the T.an hh n e t urrent uid be pre- In i a bi-n. n w ish i ,"i-ue the charge h is just '.etn cornpiettd npare ceded by a -hor .duratwn s urrent ,y . .t idie e' - N n et! tahr e o,d :,!d
@ nith i Ad and still there is n i heavy nagnitude about equal to the maumum un -un,: cents tu ttanf3 many of the points
. current in the core of the leader aNve a value of the man. discharge. The records The enginecer. the physicist, and the meteor.
The text accompansing tSs rigure would of ground gradient measurements remote ulogist can all contribute in their particular lead ene to believe t h .= *.hs charge or the from the stroke fail to reveal a current of fields to this problem. We do hope that in rew teader head in (Fi h. d bn f rawn this enagnitude. If it were present a pip this paper and the compani<.n pm:r u e have la r gely f rom *be . .ld leader -ext and mt as of great n ig utmie hould appear pruir to .nided to the kmmlede and e Anation of t ated in his hu- ai n iro n the tip d the the Ps me den nbed in Fig 5 L <uch th sto Le mechan i 2 id C.c b '.akl #n ma ler core De leader v ae ah we a is not ;4p4 hu eimen..b-erved. pr. er , of t he -p ri. a c" 'f g sps.
- -90004040 -
lt I rt n I l. l t c. .I pipe synem. whee rewn& tmrt h w as I De NOle O(r 'TOUnCing elliS and OlinflaT in s rir fo: s sr a to earth of the combined -ti ns e Dev. ices it: t{e {s(ect.ive (at{ogic be wmewhat less than o 1 ohm.
,ent ,e,1,ed te ach1 eve , -,
would be in exce<s of 15 amperes, most of hich womid c liect on the sare sr, tem. Protection of Lead-Sheathed and uuld be entirely beyc nd the capa-I _l p l_ tility of the rdvanic anode to supply. power Cal Oles O[r DUOStat, ion The result would be that the c:Tect ive. ness of the cathodic protection on the coated steel gas main would be completely Ex.ti Systems nu m .d ,d cmrosien ceuid be n pected to occur on the steel ga3 main at flaws or
- h. lidays in the pipe coating. For this SIDNEY E. TROUARD MARTIN J. MAIER reason, u ery ort is made to main-w.s ug .mc. rg o m ut tam a reas nably hi.;h reinance to rart h of the main to be protected. IIow eu r, The Anomaly Involved in the consider a well-coated steel raa main for one very cuent reason, cathodic prm Cathodic Protection of Power Cable whose remtance to greund is in the gen- tection of underground pow er cable Sheaths erat urder of mayitude of lon ohms. The sheaths is nu quite so . simple even if the open drcuit potrnt;al of this mam to a ! cad heaths are jacketed with a hi61v copper ' fate electrMe would generally die:ectric material -uch as ntruded only-OR DINARILY, inconventional tection of such the cathodic pro- be in the order of -0000 volt. To under. ethylene.
grnund structures as pipe lines, tank achieve a chance of potential of the main There is one prc,t lem indiernous to the { a be ttoms, etc., to economize in the use of of 0 % volt, and thereby bring the poten- cathodic protection of pipe type cables direct current and thereby to minimize tial to -0 M volt to a cor per culfate refer- and '.ead sheathed power cab;es which is orablem< of mthodic protection co- ence e!cctrode, which is generally con- ordinarily not enc"unt< red in protecting ordina! ion, the 4trut tures to be protected -idered as the mininmm criterion of f ull , ,th er undcry und tructures. For
. are well coated. A dettrmined etTort is cathodic prot ection of ferrous -truc- reas ns of cath,dic protteti n. *he < heath made to keep these 4tructures e!ectrically tures. it is nece%rv to cauce a direct _ _- _ _ _ _ -
free af deliberate or accidental contact current of only 15 mdliamperes to co: lect peer mm. rem mMed by the U F E in dated with <uch low rtsistance to ground 6 tris electn lyticallv . n the main. Thi e>uld Mre "ma cm?" ', e' .H n=-' ' "e " ' ".. i r e o- e n t' "" Ir '"; r *-e n' tmes as electric neut ral miems. sh41d casdy be furr ished by an ne w .sve u w uts -r m ca
~ mm.
u ms. .r .ther N rc Mer.,r und tr ic- t urnd pivanic an4 . f . nc *i. d to 'he f, j u d"* , ~ ., [,, , , $
+1 i ire < % :as u ater mains. mti .llic .< wer;
- m ain If. ' m et er. *his c'l < a ted m v'i a Rom v *' m M 't=3 't+*
- dnins. culverts or ca-ings. -teel reinfore-rnain uere a$ owed to make act : dent ai bot h .a the 1w , uns P ; Ac cr e.re I nc ,
ing mats of htH!d.mgs, etc. For cumple, electric contact with a bare buried water se. orie m t.2 0U DOHd F( .WG TSr0!G Yv!! 0l br Gd.b'h GtIwd Poi Gib!'s T* ' W R l'M1
..._._-._...~.m......._.
& EE 8 iS'N'Cill! [4.4nt' ; I i
]Oll!'! Al 1 ile I,' ntil;a
. , r, i
- v. lll [Il3tiltllC l C
DI:v0TLD '!O SCWNCI: AND Till: .NilI11ANIC ARTS Voltuna 233, Ninnber G Jtote 1.967 i I \ x - a.m:r.-=nu - -- i i f i i t i T!ic Liohining, e < Conductor i i
?
Inj u. u. color. Tiie l:lectrical l'ewirch Anociation .I l foCilllle'rflC(lil, $$lt!Yel), l$lthilli$ e I i Ill Ir oulti elIOIL I lil l15 'C t hilun.tr o[ .NeienCe, lba { l'citr: oll i { ' % Ukles "!)$'I'kplitn'll ifna%kna. tion ha< heen :u t he le t b>nt of all gre.tt -cientiti '.i-coverie<. All great .-cientist.4 g , utve in a ct . toin <en-c. bten great. artisu the nian oath na hnazination inay [ co;'ert tact s i,nl he ca"' nit toake grea! eli.-co\ eries." , 1
't*
fl !ici :.n.Z nak on: kl '. 5 nt 'ei thU aiIrihn!e U hich " L nt% Leik ko e : plain the aut-t:,nilinz succe:- achiereil in '.t t! aere t han a ilee: ale's worl; i ; eit ieb Henj:n. :n l'i.m'.:lin, t he prin t er :nn t .h pli.nu h -t. i onh! ile'. ot e to -cien.
}
- me . :n piry. Yet , rh hin that I riei pi t. "l, he clarnical .he nature of t he lightning l
.'i ei,arg nel cas e nn.nkinil the me,n, fer pron cling our hu.hling, again-t one l
..i tl.e nu <!atrnetivi force- of en t u re. !
i f t ll t he iTa lier i: -Itc 8'f t his .b .t t rfi l! IkeVlilet!
- o t ln' ([Mt*3 t i8 nt f ri li'!b t nilig pro.ect hin. 4ho ilanil ( 2) runnine.1 'he ilevelopnient of Franklin', INhtning i r...I, anil Cohen i 3 ) ih votnl a fa-c;natin2 paper in the preju. lice which follown! j
,t , oc o.i! net .on. T!n in t err -t e l icai! i - st rooulv a<lv:-e.I to a aily these tu o r i
u t1 !ii Wi! alin' [enh'!cul itill* t ri '. } t M'll I !n' u 6 it et' in.l! .i w [ri'e u-e li t t1tes flteltideiry ' of u.ine i.ect- of t he iction of 'he Ught n: *g cotuluetor :nnl to e': anti: c the -:une . ipte tion s itt t!ie h;',ht of ino,lern kn.e. leelg. I s, 1
'\ d L [
L
'"*N *' .
q, ee e. %- ,g,,,p ,% y ,, , , _ , h 90004041 ,. 1 i
i i . . i i R. fl. (loMe ' 1.it tle at tention as Titeoretical Connieleratiores action of t he ligh I Tire l' unction of a 1.iulu tnin;: Con <luetor n'\*"Y"IU"'0"'4 questum of the sp: It is a inanifestation of Innnan n enknew that a prejudice once actpiired tends had given htdo do to be retained eVen in tlle Ince of stverWl'elinill; faelual eviilenco cont radicting the bani.s on w hich it wa.s fomnled. In the leahn ol re:enco pngndice may bc In E.ngland,11. ternietl a litisci'nception. ?'nch a Idiscolleeplittu Wlo.c h has persibled for oVer F t}te polit. teal cl.amt two hundred years aint wlu.c h .is ntill w. nle<pread .s the beh.e f that a h.ghtamg any ac"~estion wI conductor has the ability, or indeed the pornose, of d. .mmat.mg sdently the result' the .luetnee eleetne chargo .m a thundercloud thus prevent.mg 'he "proketed.' bm.bling In n. tam and scicia , being .,t roc k. able t.une. IIower ' several learned i IImv did this conception originate and why has it persi ted for over two received from th< centuries? liciefly stated, Franklin wa.: led to -olving the myrteryof the IWht ning purated in a hea< discharge by recognizing mnl stres3ing it< .4imilarity with the electric epark the time. Guided which could be produced and studied in the laboratory. In his experiments he dual function of : found that a conducting body which was insulated from earth and which had the unforti.aate ! heen charned to a high vohage from Leyden jars could be sdently discharged which was still ne: , over di-tanets to everal inches by a pointed metal!!c conductor which wa3 (6), connected to earth. It wa3 this observation w hich led hun to sugge.st (4) that I.et usthen e an eartiied pointed lightning conductor might diecharge a thundercloud and nu,dern knowled thus " prevent a stroke." field of a thunde: I today. The time I
' [Iowever, in the smne letter wiitten in 17.%, viz. two years af ter the publica- h M h000 IF "lH' tion in Poor 1 ichard's Ahaanac, he runalked:"I have mentioned in reveral of proportional -
my letters, and except once, always in the allcrnelire; riz, that pointed rods IIeld "I L'00 ' erected on buildin;;s, and conuunnicating with the moist earth, would either a V"M stroke, or, if not orevented, nould cominct it, so that the buildin, hv m percent shonhl suffer no d:unage." th i ['rankhn continues to complain that, while the first ahernative was only ductors w ouc lightning flash.
"part of the u<c" of a lightning conductor, the second alternative, which was proven beyond any doubt, as soon as lightning conductors had been fitted to manner taust d i buihlings in America and ICurope, "3eenu to be totally forg a ten." This concise one function-t I -tatetaent eb>arly sununarizes Franklin's v'ews on the action of a lightning 'ind then to disc ,
i conductor.' As has been <
^IEOke OYPI GpC franklin's Warning and complaint Were, hoWever, soon [or[otten. Fullowing 1 the highly personal controvery wa;;ed by the Abb6 Lllet, and po--ibly stinm. progrenes from I elond chargo is - l lated by it, the installation of lightnin:; conductors mnde rapid progress in France. Thi< development was guided by several authoritative statetoenta of which remaiu l
'uned by the French Academy of Sciences during the pedod IMa to Isr,7. 2 la comman w I
)
' I tu a further statement m tTC,n, u nile still etni'lmt u :.; t he prevmaim n tion of a lightamg U"y" i ^ "' * * "" J rod, he refer . to the ". econd .unl principal intention et' the roth . , .. that of omJm /ing "I4 M ""*'d!""
the li;;htnin,: ' Sen (4), p. '179). l l I
- Juumal al rhe FrarAlin Indittne l
i. 90004042
l l
\
l
, 1 Tla lJgidning Concluclar }
t Little attention appears to inve been paid in these publications to the nunle of l action of the lightning conduelor; instead einphasis was placed on practical 1 aspects of the dc<ign and installation of protective sy.slems :mJ on Ihe important !
<piestion of the space protected by a lightning rod, a iinestion to which Franklin I luul given little tinmght end to which furt her reference will be m: ole later. { )
t in England, tim iniroduction of the lightning rod was initially hedevilled by ! ) the political climate of the time, ,trengthened by King George !!I's loathing of j any sugge-tion which originated hom the rebellious American "colonics." As a i renilt, t he installation of lightning conductors promered rat her slowly in j liritain tunt scientific disen.winn< on its function did not develop for a consider- ' able time. IIowever, in lSTS a " Light ning flod Conference" w as convened by .
<cvm.d learned firitish societies amt their deliberations, including opinions !
received from t he l'nited States and several European count ries, were incor-porated in a heavily documented report (5) which received wide publicity at j the time. Guided by the opinion i the many experts who were consulted, the ! dual function of a lightning rod was again stremed, thus contributing greatly to l the unfortunate persistence in the helief of a preventive action, a po +ibility which was still seriou<iy mlvocated by so eminent a scienti t as Sir Oliver Lodge i o) . l' Let u3 then con <ider the function of a lightning conductor in the light of t modern knowledge. That an earthed conductor when subjected to the electric t tield of a thundere!oud discharge., a current into the alnm-phere i, nell known i today. The time vmiation of the magnitude of this " point di-charge current" f has been frequently recorded W. Its crest value i<. as a fir-i approximation, i pmpoitional to the magnitude of the ch etrie gradient. In an average electric ; tiehl of "00 volh/centimoter under a thundercloud the current flowing through i a vertical conductor 50 feet high i< about 5 micromnperes. The charge di.wipated ! by an average lightning f! ash: is about 30 coulombs and, if the average rate of da,hing is taken to bn two flaribes pei minute, it follows t hat H.000 such con-duetors wonhl be rerptired over an area of, say, half a sepiare mile to prevent one li.;htning flash. The practical po-ibility of lightning conductors acting in this manner nmst thus he dNeonnted. and it i< clear that a lightning tod has only the i one functinn-to inurcept a lightning dischar"e before it can strike a -tructure alUi t. hen to il scharge the lightidng citrrent liarmlr%Iy to earth. As has been establi hed by the work of Schonhuul (8), the normal !ightning stroke over open ground deseleps in the form of a leader di< charge which progres3es from the cloud towards the ground. During thi< phaec ome of the cloud charge is depo.<ited along the gradually estemling leader channel, the tip of which remains at a potential with respect to carth which N only 3 lightly lower
- In ennunon with norntd practice, a Whtning duh or, more preci.-cly, a nadtiple-t roke tLoh. is maler-tno I to denote the mpu ra e of reveral in.lh idu d light ning stroki
- along cwn.
li dk the *ame diwhnrge channel. A light ning atroke doubt be under tnod v he .equence { of a sexnwar i leader followed by an upward return stroke. l
.o ;
vm '41, No. 6, June 1%7 ' lib) ) t i . ~w.., . .
.;-,.,-+.. _ _. -
90004043
J 1 [ . . l t I
- k. II Whle dum that of tlm clond charge from which it devd"In,h The inagnitinle of uns I)nring Un i potential can be inferred froni phpical reasoning, Init lhe writer prefers to ground aint um '
! think in terms of the charge m im the h ader dunund whidi can in ih,Inecil -nperiinposed d I directiv fnim n cordings of elect ni3tatic licht changes diiring tin leailer procen objects on tim l aint fn5m oscillographic records of lightning currents. ('harges from a rnn tion Further referen< of a couloinh to ahoilt ten entilombs, with an average "I 3I""ll '"" Cd" inh, can nutst he given t ' thus he 3hown to he arociated with the lower part of the dmunel of an inih- t rated to such vidual light ning st roke. through that poi ' channel ,itself, b. The distrihnt' ion of t he charge along On leader channel nmst hr affecte l by lichi to which it the po-ition of hranches which occur in the great. majority of the Innh rs ut; the first component strokes and by the electro 3tatic liehl hetween the llunulcrelonel 7.g ' and the leader channel and earth As the tip of the le:uh r nioves towants On-ground, ti e charge density near the leader tip nmst increase at a faster rate l,rankh,n hm than the charge density in the uppec parts of the channel. Ilouever, the eheet < h_scharge and d of this non nniform charge distribution can he neglected when determining the involved, impnr , electrostatic fiehl changes during the progre3 ion of the leader discharge. high voltages b. cameras and oth ' Calentation (10) shows that, on the shnplifying a"umption of a vertical the 3 park discha nnbr:mehed leader channel carrying I conlomb mn1 of a phun groutu1 rnrface. The li.:htning ci i the electric fiehl below the le:aler.varie3 according to Fig. l t a). As o.nlica t ed, its crest value i ' the tiehl gradient is highe-t vertically helow the leader and fal!< <9f rapidly in several tens or t ;
, all directions with increming distance. Figure 1(b) 3how.4 how the h,ehl increa es was reproduced i a3 the height of the leader t:p above grunnd deercases. to !ong spark g: 4 a reduced seab 1 [ -t] the tip of tle a! }~~ _.._~ . .
I ! . l l- l L _L L - a..----d - 3ithin u tra
- i. '--[~ ]~~[. .._p__ f. 1,,.,or.._l ] ,f.
M . ,!I il y.ht ! I ', /. m . _ , _ _ _ _=_
.I.,'.
l.--ll-,
. 1
---- ;c-- ! , ii.._~E4__.
-----wm=r 32- i ..p . . . . T.- 3 I.:.3,,,T m/..i,r----
..-,-.f -
l l l , ! ! I H+ . - i' -t*f, * ~'l "I il-I~ t 2 q . _ . ! . r T ' ~ - ~ ~ ~ ~ ~ F~ '
- p'-* i- -' t'! N '~l - ~ S < ! ' 4 i-- eh 'I'W
}r
* --- f }
- ; . L_ . . _i. _ q 3 ,TL=i_:;c;-.i.!:n.:' g:c:
- m mm-i i i i -
~ ! ; . ~ -- l 1 0 i!-'~C.-
ec; r- - d i. ,_ y ; .___ , __ . ; L , p J ___
-_ l. _h; --.--
;- i .__ c
.H- - l -- -
w hei e rf = t he e m . . - _ . , --l u.e,;. ,e g ,_ im,em e ,i _ i j
, p< _. _.__.-.4
. - . _ . + .
- io a
.p__.
-9r..----
. , . . _ _ ,_.. J,,, 6, = = =.d. n e
u,_ ,=_ __
--- 1 iy
- g_ .m ";f.
_ .i t-+---.--*~~- _,,g-
, if ,
{ i i . ,
- -*- .[ -
y 6-~o-
--* 7 (I {) -{ ,[, j g, -
_ .. . - l , , a _- heie yo __ _ L. I
! j! ,i ( :: i ),.- 4
. , i , : i m . _._ _
s . p%m --- , -~ _) - _7._ w,- . m r.;: :1 , m.._.3; ,d..it h the ahow I i ; ' 1I ' 'g.M_ _ _i 4: M, ,-+
~
l ; - e, . m d;+- H -t --1'( L s nT* u ~ aco
- C h I " b" . , I b. .'
,2a iwo ^
_._,_/~_ w nm , , u "A la,r 3tago, of a l'io.1. Elet trie ::railirnt below !c:uh r iluomel: 00 :ia hun tion of horizontal slemer; d') n investicition to function of leatler tie liove a grunni.1- crest value in al. s
. e .a
()4 }mrnal of Tlie I i.uiil;n Im.tvute "~ '
- 6. , a, t, L. 6. J u.e i t 0
..=.s ,
9 p D@O N D 1'= 9T3 l
# 90004044
_ . . . _ . . _ - ~ . . . . . ~ ,
. .t 1.
I i 7b f.ightnhMI f **mdrutor t I I)nring the period con 3idered so far, the current. which How.s bet w een t he t ground .uul the ilumden loud consiit utes a di-placeno.nl curirnt to w hich are f superimpo,cd the point discharue entrents which How through all corulucting l objects ou (lo gromni, such as buildings, trees, bu .lus, or even grau bl.nles, i Furt her "ference to t he-c currents is nuule later. At tin 3 ,t age, consideration l not3t be given to that instant when these upward flowing clarrents alv conceu- l-liateil to such att extent into one, or poribly 3everal points, t hat the current through that point as.-umes the n'ginte of an electric arc w hich, just as the leader channel itself, becomes relf. propagating umh r the induence of the high electric f 6 liehl to w hich it owes it3 initiation. [ i The 1.ona Spark Discharge l Franklin had already emphasized the similarity between the liahtning f discharge and the electiic 3pmk hut, becauxe of the high . pends of propagation [ involved, improved knowledge of the laboratory -park had to wait until very ' high voltages became available to produce long sparks and until fast rotating canu. ras antl other device., had been developed which primitted the progre.-s of i , tite spaik di-charge to be recordeel both photographically and o cillographically. ! l The lightning current ha< a unidirectional was e -hape which ri.-es from zero to ' its crest value in a few microsecoml, ami decay.s ntore -lowly to zero within
-everal tens or a few hundred micro econds. In the laboratory thi3 wave shape was reproduced by itapulse generators and the results of applying such voltages to long .paik gaps w ere decreed by the writer (11) and other, to repre3ent, on a redeced c:de, the di velopment of the lightning dischange at the in-tant when l t
the tip of the leader had :.pproached the ground, or a lightning conductor, to j , within "st riking distance," a term which will be detined moic preci~ely later. t ' I' This niethod of approach is not epiite correct. as can he hown by con-idering the rate of increase with time of the voltage to gionml of the tin of the leader st roke. Thi.s rate er.n be e.spre .-ed (11) by . 3' ,. - I
, a r**
,,.,'!9 n *:w' v.
s (. g a r de/dl =,l 1.S X 107 y --et---- vohn per ein. X sec. 'It) !
- y. . .. L
{ where y = the charge on the leader channel (l coulomb for a stroke of average intensity) , a a . -
)= a et ellstaltt ( 109 "< _- -
r = velocity of piopagatiotr of leader stroke ( l.5 X,JO# e m ..-ec.) / ' L rt = height of tip of leader.ahove ;rou,tul at time t. b
, , .i . oxa . ..;,. .,a, i
. . .m ,
With the above avet age values :nni for rt = 100 rii. the rate of voltage ri.-e i.s g i cen to be 2 l k\' 'ndero econd. This, then, is the order of magnit ude at. which ' . l t he voltage apphed to a lon.; spark gap -houhl incica-c if t he na chani m of the .~ ! [ j la-t st ages of a light ning .~t roke i.4 to be studied. As-unung such a laboratory i l investig:ition to he carticil out al, ay. 2 M \', t he te3t v<dta.:.c shouhl reach its T . I. erest v.due in about S5 inieroseconds. and not in about I microsecond e in the t
--3. ~ , , ..
t i
/. 4 * **
3 .,. 1[ ..g, y f
. ..r m, .n. 6. J g na .
- 455 ! i t'
/ ,
-]!' S y. , ;
, _.. .x. ,
/ ( l"
/ ',
y .- - .-.- D 9 4
- a A. M \.
A n 90004045
t t
. R. H. GuMe staiolard impulse with which mo,t, laboratory test
- on long sparks have been performed.
A nnidirectional voltage w hich rises to it cie3t in soveral lens or linnih ed - of micro ecomls is called a long-fronted ilupul e volt age in cont rast to the fa. t impul3e voltage w hich ri.ws to its cre t in about 1 inieni-ecoml and the j wave shapeid w hich has been -t andardized inti rna t ionally to repnwen t t he volt age which is imprere,l on an m erhead line comlnetor as the re-nlt of a direct light-ning stroke to llut conductor. I ong-fronted impul e voltages have assinned . iiuneased importance in ircent years ince they can he uwd conveniently to study problems arising in the inanlation co-ordination of e.h.v. lines due to switching operations in elwt rical tiansmivion -y3tems. A substantial amonut of inforniation has become available recently on the breakdown characteristics of long air gaps (12) und"r lont-fronted impnloe voltages. The.<e have been 3hown Io be nmeh moie cotoplex than the character-isties of the same gaps under short-fronted iinpulse voltages. Again, while the physical mechanism of breakdown of long gaps umler hoil-fronted impulocs 7 ' is now reasonably well understood (13), corresponding knowledge relating to long-fronted impul3es is .still ra'her rmlimentar3. However, from some work , reported by Stekolnikov (10 it cai be concholed ihat. -o f ar as the initiation of l p this breakdown is concerned (mul t.*is is the aspect in which we are here inter- i d' I
~
ested), theie is no ha-ie dilTerence hetween its mechani.-m niuler 3hort and h ng-fronted imput.-e voltages. . I Breakdown of a long air ;.;ap umler a ateep-fronted impul3e voltage is initiated ; alino t invariably at the high-voltage c!ertrode. In the ca3e of a high-voltage rod electrode, breakdown start.s by formation of corona which then leads to the development of a leader channel. The great majority of n:,, ..il !ightning dis-charges carry negative charges from the clumi to ground. They may therefore ir -irrn'ated in the laboratory hv negative high-voltage rod electrodes, the Fio. 1 .I' ground being repre-enteil by an ear thed plane electrode. In . neh a negative ro.l-plane arrangement the negative initial corona and the -ubre<pient leader
-park cover almo t the wholo gap distance before an upwant corona dischaige
- gorona and of a !
l reted to a eteep-i.- I and a counter -park are in.tiated from the earthed plane. As the.-e two !cadcr time interval or ev<
< parks meet, a return -park erptivalent to the return stroke in the lightning corona al-o develoj ;
di-"harge develop 3 aml the applied voltage collap3es. l pre-di, charges from
/ Prom St(kohiikov's work it. appears that, if the .-ame gap arranyment id meet somewhero in I -ubjected to slowly ri3ing impul<e voltages, the foregoing mechani3m may bc cotaplicated by the development, af ter the initial corona star dal the deselop. No similar infin '
l mer t of a short leader spark from the high-voltage rod electrode, of a mid-gap rod- rod gap under i I breakdown proces, which progi .res simultaneon ly mnl in opposite direction.s C"H"'s available, it i i towards the high-voltage leader channel and t he earthed plane. The ocentrence in the case of the ro 1 j of an npwant spmk di3 charge from the planc appears to be similar to that maler l the influence of a -teep-fronted impul-e. The 17nol %;te g 1 i li, now, a rod electrode is mounted voi tically on t he cat t hed plane, con- llaving thus for. I verting the rod plane into a rod-roil configuration, the initial ocentrence of !"uithnaie to a+nnu
) g,,tl f lL4 Of DtC r(3fth 1Q lthtl[ldge *s 6 8be Ili MJ s
^
***"N****- ***'**='d*'" ** Mu' M
- We asp - + =w ee**a-e. w e % w a,,gh y. -e-.- -s- =* * * - ' aa'm *%W_ -_r****'-
'* W"
90004046
1 l 1 l Tlw I.;:thinin:t t %r Ina tu,e i
- , . _ _ . _ . . _.. ~ . . . . . -
i t i t t t i 1 i i s
\ i i
t 4 f l l l
?
I I
, -l k I l '
- s i
k-.~.... . ....u.x.-.-m., l't:i 2. l.i;;!:t:iing stroke to e liinutoy pot ..how ing niectin; In. int In't a ren downw aril lentler
;uel upw.iril stie.uner.
e \ coriiint tilul of a leader ilischarge at the negative higln. voltage elect rode :uth-
)
jecteil tii a -terp fronted ilupttl e Volta e is unch.ingoil. If anvever, af ter a brief tium int"t val or even shutill.titeista. ly w it h t he on-ci of t he high-volt: ige corotta, I. corona al-o develops fioni the tip of the earthe d roil electrode. and the two , pre-ili.-cliarges frota the high. voltage electrode and front flie earthed point
!neet intitewhere in the innl-region of the gap.
.No sitnihtr infornnition is available ott the inechanisnt of bre.ikdow n of a rod-roel gap under alouly ri.-ing impul,o voltates. I'ntil -uch infonnation be.
entne3 available, it unt t he a3stitucil that the basic nieehani-ut i, ullehallgetl as in t he ca.se of the nni plai:e electiode sy3 tent. Tise l'ineal .%t o::e of ou 1.i;;In toring .%trolce to Grotortel . 6 l[.tvilig l!LU4 [tirlheil a llielitre i)[ llti lli'e.tkillin n f >f a linici air Zap, it 3eelnS
!egiinuate to a.einne Ihat the final stage in the ilevelopinent of a !inht ning st roke i u . n . ~. ,- - s .m
> M.
i i^ 1 t , ~.- - . . . , . , .. . . . , , _ . - . - . - - - - - - - - 90004047
f i R. II. Uvble to earth or to an eat thed ohjt et is delerniined by t he -:nne phernauena, thtid feet) innler innat e revetIing once again to Franklitti arginnent about Ihe siinilirily helwecn t he chgsing gh.gg,jg glg, lighttiitig sli chnige and t he elect rie 3 pat k. Pluitour.iphic evidence to suppot t u hich g [j hinjg g t his aramnent is dillicult to obtain' inainly because of t he remote po3rihihty of h ader cinae windd
-ecuring a photograph wit h a rotating camera of a clo+ light ning discharge and becau-e of the high 3 peed, of propagalion ovet shoii lenalh3 of uch a di.-charge.
The writer is all t he more happy to repnnlure the still photograph t Fig. 2) er a T^"' k I a l'er clo-e di charge kimlly placed at his di posal by .Tir. Ve.-anteiii. 4 The pimtogralih -hows the lower emt of a light ning di-charge lo t he chimney C,""b""d"" pot of a buihling in llelsinki taken at a distance of nhout 213 feet.. The dis. I charge Inu t cle:uly lutve Imen of low severity as indicated by the -mall atuount -- ~~ of ludation on the original photographic negative aint by the absence of damage Hod -rod to the chimnev pot apart from an acennudation of soul. ()ver tuo3l of its length M".1- rod the photograph 3hons the honinous core of the lightning chattnel, but near the "" "d centre this channel is clearly divided into four di3 tinct parts. From an analogy liod-plane of sitailar photographs of long spark discharges it is . uggested that it is this Hod -plane l point at w hich the dowinvant nmving leader 3troke is met by the short upward ""d - Ph ""' Hod-plane leader d.i-charge from the chimney pot. i Rod -rod A few other photographs have been publishnt .-howing the occurrence of ""d P "" i upwan! -tre:uners from ohjeels about to b" struck
- so that it may he accepted - --
l that a lightning leader stroke proyres-es in the form of a self. propagating tiis-
! chatgo from a charge center in the cloud towants ground until the gr:ulient at a in the upper p i-i point on t he ground surface has increa.eloulliciently to can-e an upward streamer h>;"!her which h..
of t he type rhown in Fig. 2, to he initiated. The que3rion then ari-es a, to the sokaps ii-ing t critical vahic of this gradient. ynp a.., . g - he reg w Ilea-ons have been given to *how that the gradient at the ground surface 50 per een. be!cw a downcoming lightning le:oler stroke increa.es at a rate which is similar can he unumi, to that occuriiiig in an impulse voltage w hich risc., to it crest wit'a in a period of a feu tens to a hundred microecondi The breakdown voltages of long air gaps under 4ach long-frontnl impulse voltages have been determined experi. . f"r cod iod ga; mental ly dming the la t few year <. A critical examination il2) of the re-ults mere: nu dWance obtained by ditrerent investigators hows that, while average breakch.wn volt- ch' eaa he accepte' ages and their statistical deviation can he established for positive impulse down gradient deen vol t a gm. the available data for negative impul-es ce too meager and too in- fixum "f about 7 kV con-i. tent to permit any generalization. Pimtive light nin. The r"ason for thi.s polarity elTect is the inuch higher negative, as compared einu ge in temperme with the po itive, breakdown voltage as a result of w hich an impuhe <,enerator . inf ormation on the producing a long-fronted voltage of 2 .\[V is capable of causing a brea!alon n of p,q(n,Med impuk up to 6 meters (20 feet) under po<itive impul e , but only up to 3 meters (10 aver em of a 66-2 pror. Merger's emnpanion paper ' pw m hotihnh.s shou how dec ,Wu'He- have ro 1 -n.d mal rod ph-heen overcome. clea-e w it h increasi' I
? ' ' , It'f [halahCO, I5M. II, l8 q0 M I, ih t he eutnpanbW lI kpPr !'y N.IK'lef. gljlfg.g pgjpg, })pg gppgg g j 158 u,n<a.a nc n.muioi-nnn. ya y y s , y s
~ . _ , . , , , _ , . , , _ .,c.w.,-.__ _ . +~-c ~. - - . - - - - * * * ' * . - - - , - .
l l 90004048 1 l t I l
1
- . +. - - _ - -. m . - . - .- .
I i 7'Is 1.ightnimj Cominctor i feet) ofuler negative i:npol3es. This ohnervation leads to the it teresting con- I cht-ion that, in the rare ca. e of a posilive lightning di<charue, l'ie di tance over < u bich a lightning conductor wonhl exert an at tractive HTect on a downcoming le:uler stroke n onhl he greater than for a negative stroke. f
'l '
Tann.n L .in l'er l'a at Suita hing tmp olse Hantdurcu (lor ih nts of I.ung Sp u k (laps i i ; i l)i.-tance in ineters ('on ti;ciration Pol.u ity 1 2 3 6 .i I!cf. llod-rod - 6.7 6.5 5.9 (15) ! I t od-roil - 4.0 7.2 (10) , llod-rod - 7.0 6.0 (18) i liod-plane - 10.1 7.4 6.1 (15) f I l 1lod -plane - 9.0 (16) flod p!:uie 1I.3 9.0 f (17) l 11o1 plane - 8.I T1 (18) ! I!od-rod 4- 6.0 3.2 1.5 4.0 3.6 (12) llod-plane + 1.5 3.7 3.3 3.0 2.S (12) { i _ _ . _ . _ . _ _ _ _ - _ _ . . _ _ _ . - -_ _ _ .-- _ _ __ _ _ _ _ - _ . j i In t he opper part of Table I the .30 p"r cent breakdown voltages are col!ceted to ;et her which have been obta:ned by di!Yerent anthors with negative impul-e voltoues ri-ing to their crest within 100 to 200 microscenials. Dividing thesc . voltages by the breakdown di. stances produces the vahles li-ted and these can I he iegarded as the critical gr:alients capable of producing gap breakdown in ! 30 per cent of all voltage applications. The information contained in this table ! can he sununarized a4 follows for breakdown distances frmo l to I meters. l 1 lr rod-rod gap
- the critical breakdown gradient decrease, only !Htle with inere, ing distance and for gaps of the order of I meters a value of ahont 0 kV/
cm. can he accepteil a< repre-entative. For rod- plane gaps the critical break-dow n gradient decreases with increasing di tance and, for 3 meter 4, reaches a IIgilr( t tf ahoilt I kU,'em. Po,itive lichtning flamhe< eon-tirote only about 20 per cent of all carth dis-char:c-in temperate region and even less in tropical regions. The corresponding information on the critical breakdown er:ulient of long gaps under po itive I me fionn d impul-e voitages is mlded to Table I. The value< ynoted constitute wera2e- of a fairly lar:e body of esperino ntal ilata. It appear, t hat. hath for , ; ru l ro I and rod- plane gaps. the critical breakdow n radient continues to de- ' { ema-e with increwing distances e.ithin the ranse investigate 1 :not that the ditference het ween thc3e two contigorations i.s greater than for negative polarity. ,
% r .U. la. c,, jw.e ywd 0 l
l 90004049 i
. .I
i l R,li.GoMe Win n applying these experimcatal data to the nat ural lightning di-rinirge, in re is inodified ari' r everal factors ought to be con 4ideied. l'irst and furotuost, any iptantitaiive State l'uilding or of extra]Hilation front the cale of laimratory terls oli gapw nn asuring a few incters I'fl"M M"UhlNnhI to the reale of ihn lightinng di3 charge inn 31 alu ays he subject to certain inental of a downwani lt"uf reservations. Portinintely, an uncertainly wit h re peel to t he critical breakdown that t he !cacer A gradient within the fairly narnnv limits invoived enn he rhmvn (10) not -eriously doctor leading to . to alTeet the conclusions reached. o t he pre.<ent contt The secoml niajor con-ideration involves the geometrical representation of The Strikin g Di.<t<. a lightning discharge to ground by a rod-plane configuration aml of a lightning In order to deter d.t. charge to a h.g htning rod by a rod rod arrangement. Cons.ub ring b.rst the
. . opu nni ih.3 charge.is l.ightnittg leader channel, .t< i tip b urroutided by a corona covelope and the same ture, it appears rean occurs at a h.igh-voltage rod c!cetrode no that, electro-lat.icallv speak.mg, the t hat t he en. .tical bret two arrangements can he legarded as .mular, the ground configun There is little objection to representing a single vertical lightning rod by an "?ative Ughtning c earthed rod elect rode. On the other hand, open grmuul cannot, strictly speaking, \\ ith these values th he represented by a plane electrode. The earth's surface is nornenly covered by upwant streamer is '
sinall growth, gran bla les or stones .uul every such projection is capable of The rc<uits of thi4 c: producing a paint discharge, as beautifully demonst rated by Schontatul's .\s indicated by t "
! measurement of the point di3 charge'eurrent produced by a 3 mall tree. Similarly taboratory investigations of long rod-platu' gaps have shown that, particularly -
i for a positive plane which corre ponds to the nonnal condition of a negative l lightning di3 charge. roughnew of the earth's plane or 3light projections mate- *" rially affect the breakdown mechaniem so as to make it more similar to that of , the rod-rod arrangement. :
.\ third point worth noting is that the present discu nion i< confined to the case of a vertical unbranched leader channel. Mo<t lightning di.-charges deviate l hy a gteater or les cr degree from the vertical and both this fact and the oe- i l currence of long branches from the main leader channel mu3t canoe a distortion ,,
of the electric field about a lightning conductor. The smne applies if the lightning ! conductor is not placed on a uniform plane but on a buiMing m on sloping 2"- grou"d. These (netors mu t clearly atTect the result 3 to be derived from labora-tory nvestigations of the type con <idered here.
, .\ir pre 33nre and rain fortunately exert only a very slight effect on the break-down voltage < of long air gaps.
From the foregoing considerations it can be conchuled that, as a lh-t up- 1 m. a. %: proXimation, it is legktimate to visHa!!N the normal natural lightning lemler
-troke as a 3 elf-propagating discharge which progn*-es tmvards ground guided poih on the gromul by the local ticld di tribution in front of the leader tip but unaffected by any nology which was sig featuros on the gronmi until the critical breakdown st rengt h of t he remaining i< a function of the i: i di<tance fn m gromul is reached. When ihis stage has been reached, an upwant , y g,,, g, l
{ . 3 streamer discharge is initiated mal t he 'cader -f roke is diverted tow ards it. a,a.h 3 tr stmsure
-l The only condition in which the " normal" light ning di.-charge di<cus. sed Si re uler i., refe rrat to !
i.
).sullMk ul nW l't .l d i t si l s s alil tit C 85. D i. \48 fr, fullt* eb[
. _ . , . . . . _ . . . . _ . . , _ . - - - - - - , . ~ . _ . _ - _ . -
90004050
. . _ , ,-. . . . . . . _ . . ~ . . ~ . . _
Tlic l.ightnin:t Gnuluu tur here i modificit ari-es in t.he cue of a st ruct ure of the in inhl of t he IUnpiro , . State linikling or of such tall towers e us d in iterner's i 19) li.:,hining observa. lory on .\lt allit San Salvat ore. For. l Dit't tile 4 Inf Sin'il licithl, ille ni8rntal pnWes s , of a slownwani leader troke followed by an upwanl reinrn st n.ke is reversed in i that the leader -troke i- freipiently initiateel at the tip of t he tall lightning con. duelor leading to :tn upu ant leader 3troke followed by a p3eudo.rcturn st roke. In the pre ent context this type of discharge will not he con-kiered any fuither. i l 1 Ilue Strilaina llistance of a Li:.;in tning Strolce In order to detennine the height above grounil of the leailer tip at which an i upwant di-charge i< liable to occur from the grotual or from an c:n thed .<true. t nre, it appear, remonable to uggest, on the bei, of the data given in Table I, I that the critical breakdown gnatient i<. a a tirrt approxiination, indepeintent of j the groutul configuration aml that its value i.s of the onter of 5 kV/cin. for the l I negative lightning discharge and 3 kV/cm. for the po-itive lightning stroke.
)
With the-e values the height of the tip of the leader can he calculated uhen an upward treamer i. liable to be initiated from ground or from an earthed object. ) The results of thi calculation are hown in Fig. 3. i l
.b hulicateil by the curves in Fig. 3, the height of the leader tip at which the I t
i i f a s.~ ,- ,/
/ e
; r:c~ , e umvs poor t y E
/
*e ' ,/
k
,[ , -(tw poicr.ir 2 ,,c_ e /
un ; >
/
I - - l mL y/ / [ j
^//
o w "h lo a lw >> a 4 L:y ts r.) C.,rr e nt Amp utud e , h A I ten..t -tnkin,: dist ano-i or nea;atne an.i po-iris e lightning etrokes. . point om ihe ;roomt to be -trnek i, actenninca or, io n-e ihe de-ciiptive termi. nolo 2y which was -ignificanily already u-ed by Franklin, the "-triking di tance" , i, a function of the inten-ity of the en-ning lightning -t roke This conclusion is ( ,
) I i At the pre-ent st.ttr .d know h,1v no indv ation c:tn be gn en of t he ennditu.n.s uheh-r ,
w hich 4 "t.til" st rutt ure is ca p.shh. of c.in.-ing upw a rel L -der -troke, to des i lop. l!% n er, the remier i4 rrierred to 15crger' ioinp;tnion p:tper in w hb h t hi4 ' tunnon i .hscus. ed. : l l f P e 10l i v.a m. u ,.. % c,a f i ; i l _ .._ J 90004051
. -. . . . _ . _ _-._..._.__._._.____.m. . . ~ . . . . . . . - . - . . . - . . - . - - - . . . . - . .
l' , i l, R,ll.tiolde 4 e original phot op,raph t rack of t he lif.,hinit
} at which the downt charge front t h e 1.'
"st riking dist:-
l above : uni in !o..
- l. was not inuch inore of moderate inten<it ot her photograph.s . l d
{ strokes are availabh y indicated in Fi;r. 3. s f Several research l
; calcuhting the stri! l
,,.,s., ,i the scope of this pa I
, J <
inves tigations. L' ,<n s y ln ~ ~, , :(,
~ '
f m .: \
\ ?
1 ? Th e . It tractive Rar { surel the Space Pro [ The sttiking dist j' tip of the leader E struck is determine. hy Ihe geomet rical lc. --_ - ._ _ _ _ . ....m_ _ . - - - a aroke to open - Fnn I. ii.,htning stroke to thinumy. a!3o innlicste stroke an.' pretty ohvh u, if we consider that the electric gradient under a leader channel the strok.- is a functir.i of the charge on the leader channel (as shown by Eq.1) and that "t he attr. this in turr ;..oportional to the maplitude of the current in the return 3troke e,unluctor att rac-to which it gives ri. e. Thi< striking di.-tance is greater for the less frequent constant but which po,itive stroke than for the normal negative lightning di-charge While this distance may amount to 200 m or .stWhtly more for the rare, very intense light- Once a",ain it is ning !! ashes, it amounts to no more than about 30 metcrs for an average negative Franklin. la a lette lightning stroke of 20 kA. "The di.-tance at u suddeniv, strikine th l{ everting to (110 photogt'aph *hoWn ill Fig. 2, the height above the chimney :.o highiv cinu ged, i.s struck of the . ubdivi-ion in the lightning channel i.4 about 9 m amt if it is de- ,;, , a,ld form of & duced front evidence provided by t he long laboratory spark di.-charge that t his Thi, diaanct, whmi , meeting point of the downward and upward discharges occurs at a height of g,. q.j,yj gj,qqm.c, y about 60 per cent of the total striking distance, this latter may be estimated to y.ill bE nnute." have been about 15 m. Front Fig. 3 this woubl indicate a light ning current of about 0 kA, ri: a weak discharge, a concin ion u hich had already been reached The problem of from other evidence. fnn her que.-tion of Ih ifele a cl Par di linel j Figure 1 hows a light ning st rike to an .u. foot hk.;h fact ory chinmey or, to be v itical Ibbtning ro i Inore preci-e, as can he clearly seen with the aid of a mag!nfyitig glvs from the by .,uch a condnetor.
-N Journa! .A Lc l rakhn luotate y,,i f o i, n ,, y,gg gw I.
. , , , _ , , , , _ , , _ ,,. .- .-,~~--~---------o~------
T 90004052
-l l
l TIw I.h;Iotrob>g M lact.or original photographic negative, to it s light ning conductor. The sharp hend in t he i tiack of Ihe lightning charmel can, in thi, case, he taken ta conslit ute ihe point ' at u hich the downcoming lemler shoke was first attracted by ihe upwant ili-.- . charge from the lightning coiuluetor no t hat this distance wouhl rignify the "ntriking di-tance." llaving reganl to the similai light intensities of the track above :uul below the knee point, it can he deduced that tho . triking di tance , was not. nmeh nuire than about 30 meters, thus once again indicating a stroke of moderate inten3ity, as also ouggested by the small muount of halation. .\lany ot her photographs showing sinular, sudden changes in the tracks of lightning t roke . are available supporting ihe onler of magnit ude of t he st riking distances j indicated in Fig. 3. l Seve al research workers 00 22) have followed the author's approach in c;dcolating the -triking distance, of lightning strokes. However, it is beyond the . scope of this paper to discu s the varying assmnptions or results of these , l inves tigations. l l The .f ttractice Hunae of a Ligh tnin: Corulactor ' orvi the .ijmee l'rotreted by it The striking distances plotted in Fig. :t were calculated as the height of the . tip of the leader channel above ground at the in3 tant when the point to be j
-t ruck is detrimined. If, a3 argued earlier, this . striking di.-tance is unaffected I h3 the geoniettical configuration of the earth electrode, viz. if it is the -ame for a .~troke to apen ground as for a -troke to a vertical lightning conductor, then it aho indicales the maximum horizontal distance between the tip of a leader otroke and a !ightning conductor over which the latter is capable of attracting the -troke to it elf. This di. tance may be called "the attractive dietance" or "t ha attractive range." It thus follows from Fig. 3 that a single vertical lightning ;
coaductor attracts to it. elf light ning stroken over a di3tance which is not a , constant but which increa.ses with increasing inten.-ity of the lightning discharge. [ l l Ouee again it is interewting to note that this hlea had already occuned to [ franklin. In a letter written in September l~6~ from Paris he states (23):
'The distance at which a body charged with thi tiuid will di-charge itself !
-uddenly, striking through the air into another body that is not charged, or not ,
m highly charged. is different according to the ciuantity of t'.e tlnid, t he dimen. ;
< ion, and form of the bodies themselves, and the state d the air between them. -
This di. tance, whatever it happens to be between any two bodim is called their l Alff[.[D// (((.4UN'c, as till they coine Wit llin that di-tance of each other, no stnke ,' w di be made." t i The proldem of the attractive range of a lightning conducts.r rai-e4 the i fm ther ilue-tion of t he space over which such a coinluctor wiil protect a huihling.
!!ere a clear distinction needs to be made bi tween tho ar t cartive range of a vertical lightning rod w hich has been di3eus3ed m far aml the -pace pn,tected by -uch a conductor, a de tinction which is frespiently overlooked. The altroclire >
1' 1 i i e pen . I v..l. .'3 8, h A. f une P)r4 'Il h ) 5 =-- ms.. m _ . -m,--e.~, n..--. - - - . mm - 90004053
- - - . . . - . ..... . . . . . . . . ~ ~ . . . . - . . . . . . . -
y . - . . . . i l R. H. tloide i -eenl4 lit tief llr in f llp ' of Science 4 (25) u hii I In that edit ion, l k hghtning condin lor
. .t .-
{ -l> ace ahingt it, lhe rn A'l % L D % -, f ., h . on (26), "the acqui s
; .T
"'""-%- & hg. , ** ' ' "' ' O */ , % iwult s the creeth
[ Q
'%- - l _f.1j%( . N, h:nidings so as to inci
.b an interesied oh l !.I J M, %'4 ( Ty .J 7 b i~~~~f k . f) \
4 .%
'i
',fff.{l
.* ?;
j] ~
- . '~,' *
( 3 i
"*'"D
" a4 I"d hF I?I X 0-utuloubtedly ju-p' . et . ,~d If"""h \P"dC"lF "I
, k,# L, -Q _f. $1.-,LSL,j.d Y
vdalinipodance of f. Fm.1 Tall lightning rods on Gernan ca-tie. protected by a lightni range of a lightning rod de.-cribes the di3tance over which a single vertical lightning rod of given height standing on an iuidi-turbed plane can be expected In subsertuent edi j attractive di-tance t, to at tract a lightning leader . stroke to itocif. The space prob efed by such a con. front '.' to G5 and i; l ductor, on the other hand, .-houhl define the -pace over which a light ning con. Preece (27) it was (, duetor erected on a building of given dintension- can be relicil upon to protect i the 19th century on l the building finnt 14ing struck. in a thunder 3torat the stuface of .such a building is wet and, just like the r--- --- . surface of :. tree, is capable of carrying currents w hich inay be .-nflicient to
.-upport point direharges frotu corners of the roof .-tructure. To what extent TJ t this etTect can reduce the range over which a lightning conductor is capable of l diverting a lightning leader 3troke to it-cif is unknown. Fortunately, this ques- ,
tion is of ininor practical importance :ince modern lightning protection insists o an litting lightning coutlttetors along rideos or parapets of roofs 30 that all O j l /- protuinent corners are etTectivelv protected. _t lit exi3 ting publications the e.sential difference between the attractive j g rance and the space protected seem.' to have been widely overlooked atul it is l f on this understanding that historical views about the space protected by a l lightning conductor inay now be ex:unined. Beyond the 3tatrinent just quoted, l l Franklin 4 does not seem to have made :uiy pronotu:ccment on this topie. The -- - - - ;co m _ _ _ _ l staternent itself appears to have been completely overlooked by later investi-gators, and the fir 3t referettee to the protective range of a lightning conductor ,, l.ro. 6. /.ones of pron e
- Indirect evidence on Franklin's view regarding ihe < pace protected be a vertical lightning .Iltr ro 1 may be deduced from the f.unous case of the powder nwgazine at l'udh et. <241, the hght. Ju(
In: ning protection of whh h wn= de-igned by a conunittee of four scientists, one of whom wa.4 f . F( Franklin. .\ corner of this buihhng wn- .lamaged by li;htning. The ratio ..f the horirnotal iz u di t wee between t his point an I the lightnint rod and the hei::ht of the tip of t!.e rod ahmn t;;i-the pr. int .druck was 1.tti to I, -o that this " protective ratin ' wa4 app rently reganled as Fu ade.putn by Franklin. It may be noicd that one awdn the light il.unage in licate* a com. l[,y paratively weak hghtning ii.winrge. vol Mt, L r, Ne my d Journal af The i tanklin Inentute [ ._-.....s- ..
~ - -..
90004054
< - . . . . + . . - . _ _ . . . .
. ~ . . ~ _ , ~ . . . . . . . - - . . . . .... .-..... . - . -
t I t The UyMning s '..ne'ar for f seeins to occur in ihe"in t ructions sur le.s Parat onnerres of Ihe French .\cadeniv of .4eience4 (25) which were fir-l issu"d in !S2:1. 1 i in that, edition, which was prescuted by (lay-Lussac, il was stat ed ".\ lichtning conductor protect.* elicetively ag tinst a lightning stroke a circular ! pace about it, the r:nlius of which is talte its height." .\n .-tated by it. .\ inter-
.-on (26), "the aciptie.-cence in this suppo ett absohite forruola h:ul for one of it s re-ull* the erection of inonstrously huge rods nuule to tower high above ,
huihlings .-o as to inerca. e the liehl of protection to tlie large3t pos3ible extent."
.\ 4 an interested ob.-crver can .-till notice on the continent of 1.'urope atul as ,
exemphiied by Fig .1, .\ndern in's complaint about monstrou-ly high conductors ; w ts uruloubtedly juslihed. Yet, be thin as it may, great eiedit accrues to the French .\eadeiny of .<ciences for having fir <t drawn oilicial attention to the , vitalimportance of fortnulating acceptable rule.* on the space which is cifectively protected by a hghtning rod. i In .-ubse<iuent etlitnins of the French Instructions ti.e oriuinal ratio of the , attractive di-tance to the height of a vertical lightning conduttor was reduced j from 2 to 1.7.3 :unt in t he first th-itish pronouncement on t his <ubject made by i i Preece (27) it was further rerluced to unity. The various -uzge.stions inade in ihe !!)th cent ury on this imprirtant ratio are illustrated in Fig. G. f [ _ . . ._ _. . _ , ( r _ _ _ .. _ _ _ _C _ _ _ _ _ _ _ 0_ A P ,' V 0_ _m u_ _r ir_s _ _ _ _. _ _ ._ _ _ _ K i A TJ L _, i fill $NL / l \ i / \ N!fN N ! f 1r
/
/
,/
/['
IIo
~!$
['
, _ _ 'N
'N f
)
t H G f;. C [ B g F ,,, l l 1 I l a l l l l I i i l i l vl1 5 via 1 l i l l
' I e i l l l~ ; ' '
l i i i l i i i _ _ _ _ co m . . . _ _ _ m m . . . _ S._.50 m g --50 m -- p---175 m -Nrr.{ i ! l . ) l P Um. 6 'ones of protection by vertical H.;htning rod (after O. J. l.od.te) (6). JllC K cs liinter (lav f.u .-ae 18Z1 II.\C cone I)e Fonveille IS7 l i).\li cone Pari < Conuni. vion 187"> LF(! AI c,s linder chnpenan iS75 . I'.\(i cone .\d.inu IMt } g i)llIP o linder h31 ot hesix F.\( ! -pes hl cone l'reece LAsl l 11.\1 core .\fil ens Ub) i a m. .h t,. bn : ev,r f I 9
. . -- -ee, saammy. . e e- es.e*=, .ar*.-e.es e .w== ew=om = #-w e ase**==.* ==**
*9"** . .
90004055
-....,_.._.<....~.s . - , . . . . . .
s
, R. II. (Mle Credit is frequently given to Walter (28) for inaking the firft attempt at, 1) If it is :u cepi dMermining the protective range of a lightning cominctor by collecting factual function of t he elect inforntaD.,n on points struck by lightning in the iminediate vicinity of elmrch tractive effect of a I steeples equipped wit h lightning conductors. This overbioks the fact that about along the leader ch hfly years earlier a ncienti-t such as Oliver Lodge (0) or a practicing engimer like dillicult to se luiw It. Anderson (26) had already drmrn attention to many instances uh " N!.t-ning h;ul 3 truck a building within the " protected" zone suggested by various 2) The hirakdo.
authorities aml h:ul reached the conclu3 ion that, to speak of a fixed space of greatly alfected by t protection, was "inmhuimible." clearly established I, this abo applies to Ibewever, once again all this earlier work was forgotten when the introduc. writer i< at pre.<ent u tion of the high-voltage imind3e generator nmde it po+ible to apply to a variety a definite effective in of tent objects very high voltages which were believed to ninmlate the wave shape of that produced hv a lightning stroke. This development initiated the
?) IInvn.n' re~mr.
period of laboratory model tests to determine experimentally the attractive unpulse te.-ts on lon"" range of a lightmng conductor, thus reverting unconsciously to the unhappy tents have freiptenth - h.eation can be foum, llenjamin W,ilson (29) who opposed the 1,rankh.a lightning rod and who tried to prove h.is pomt by tests carried out w.th i an enormous battery of I,eyden j. ars by a posit.ive rod cle-in the Pantheon, a larac buihling in London's Oxford Street. To munmarize, ti F. W. Peck (30), who wa.s the first to mulertake -y tematic model tests, the author to the c. found that the attractive range depended on the height above a grouml plane de in cla mim th lightning condu tor of t he high-voltage electrode chosen to represent the height of the thmulercloud. Taking thi4 height as a thousand feet, he found a " protective ratio" n hich varied rme of a lidhio g;g .,. g. .g. between 2 amt 1. Peek's reputation was ruch that this figure wa.s incorporated Fig. 3 $re ;.ccm in the l032 edition of a U.S. Code (31) in which it was retained until IO4"> > , attracter when it was replaced by a ratio of unity for important cases aml of up to 2 l for less important cases. even thi-part of l Once it had been establi3hed that a lightning leader 3troke is " unaware" of any feature on the ground until it has come within striking di. stance, model I,ru c t u.rol .lspects . tests were carried out with the high-voltage electrode sinudating the tip of the leader channel at a height above the ground plane which wa4 grmlually reduced A lightning conj to the height of the grounded rod electrode <imulating the lightning conductor. ductors, the d<ntn e Many such test 4eries have since been pedormed (32) ami the result., have { be brielly examined. been applied to determining the attractive effect exerted by a lightning con-ductor. The Roof Conels Even with tlu.s latest, refinement, there ew.t m the author'.s view several Franklin'4 public was made .m 1.,oor Ib obj.ections against the acerptance or. results from model tests to a quantitative determination of the attractive range of a lightning conductor although thev (34), Franklin euw c'ght feet above t he can be or, con 3nlerable value for comparative inve*tigations (33), such as the A n. e b h.ng etTect of a ground w. ire wit h rc~pect to the pha-e cominctor.- of a 3 tater uif the house and a mhldling wire t ran mi3, ion system. A part t rom t he rate ut. voltage rise win.c h -houhl m. the future be used for 3nch te.-t3 and which has been disenmed earlier, the folhming [n we,,jgg pr3g major objections may be advanced: n m ,t be remendierei Mfi Janu .,r rw rnatim insotute ui .S, w n, yon, cw
- - ~ - - - - - " ~~
~__ _
, , , . , _ . . _ - . , - ~ --
~
90004056 i l
.......m...-.,. .-.,a.,-m
. . . . '~... . TI.c l.Ql,tning Omulmier l$ If it is accepted that the striking diwtance of a lh.thining discharge is a fitnetion of the electrie gradient hetueen its tip alid t he ;:n dinil, t hen t he at. t ractive effeel of a light ning contlnetor ino-t vary wit h the rharge ileponited dong the leader cluumel :nnl thus with t he inten-ity of t he sli3rhan ge. It i* dhlicult to see how this feature e:Ut he ,inmlated in a hihorntory test.
'.' ) The breakdow n mechanism of a long 3paik gap in t he hiboratory is matly alTeeled by the seties teAlance in the di-cliarge cirettit. This has been i n arly ( 4ahli-hed by ic3ts with iast-ri3ing impul4 voltages. To w hat extent this al o apphe3 t o (lic luot e slowly rising iinpill-e volta;/,es 34gge3ted by t he n riter i4 at present unknown. On the other hand,it i< dehalable to uhat extent a definite elfective intpedance can he alltibuted to a lie,hining leader chaimel.
;l) IIaving tegard to the very high negative impul-e vohages reiptired for impulse tests on long spark gap, and the great dispersion of the results, model
! cst- have frerpiently been made with positive impul-e voltages and no justi.
fication can he found for trying to represent the normal negative leader -troke hy a po-itive rod electrode. . I-Tii -inuntariw. theit, t he phy -ieal voitriderativin3 outlined in thi4 paper lead t he author to the coia lusion that titiver hodge and 1 ichani .\mler.'on were ! right in claiming that acceptance of a fixed vahie fier the area protected by a [
!ightning conductor i< unju3tified. Exprered more po-itively, the attractive e I
rrnge of a light ning comhmtor -luulld he regarded as a tal!-tical eptantity de-pending printarily on th ' seventy of the lightning -ttoke, if the curve, giten in Fig. I at e accepted a.: a guidi . a lightning -(toke of average inten-ity would he ' d 3 ' atirseted over a distance of ahont twice the height of the conductor. Irowever, m , [ even this distance might he reduced by an unknown amount if any unprotected part of a huihling was of such a shape and in such a position as to be e tpable of l Illitialitig lIn upward *Ireatnel' discharge. i l'enetical . l3pects of tite f.i;.:Istning l'rotection of Structures .
.\ light ning conductor y.-tem compri -es three main pat t , the roof con- !
I ductors. the down conductors and the enithing arrangetaent; the-e will now l he briedy exan.ined.
\
The Hoof Conductor Syntem Franklin's public introduction of the lightning condnetor, it will be icealled,
.<as m..de in Poor l!! chard' .\hnanae for l-'.11. In this briefe-t of -pecir. cations
< .3-1,, Franklin -nggest 3 to ptovide "a sinali iron rod" nloch "may he 4ix or eight feet ahm e the highe-t point of the building.
- A few +ntence4 htter he i
'tatew. "l( the hou-e of harn he long, there may be a rod or point at each end and a mid.lling wire along t he riik',e fiom one to the other. '
! l ht are--ing Franklin's idea 3 about the hghining protection of -tructurc<, it ; j unt3t he temembered that he was largely concerned with -inall duellings and !
! )
i a' . va m..w o.J iu , F e - . .c ,= . _ . , , . , . . , - . , , , , , . .
- c. **~<..ap,., , .
90004057
t ? I It.II,Unide i agricult ural buihling*, netuy of thent inaile of wood, w hich w ere. ;uul till are, juodored by the I,e fretjuenlly destroyed by lires started by lightning. llourver, even within this catlier, a lightning i re.-t ricted range, he alre:uly recornnternled t he t wo tuost itupot t ant p:o ls of t he occui t ence of a hn!. this con tructioi roof coiuluelor sy3tein, ri:. one or tuore vertical linial- :oni a horizonf al roof s coinlue t or. houn hv the a Ict uat ively, -a no. f
.\lthough 1*.rankl.ut look a keen m. ierest ut all ca es of h.uht ning stinkes to f' conductors for Inin binldings ulu.c h had been protected by h.is h.g htning rods, :uul although he i
late-t edition of ano utdized the experience -o gained .ni d.ren -mg .-uch problems a . the neces.,ary i cro i rection of a lightning conductor or the tisk of .-ide flashing, the esperience .\s -ugge ted e:u
. emblcil during los lifetime was es crely lignited, it i , therefore, all the mme attract to itself evi notewolthy that already in the first, in-tritetion for the protection of a "long ' the coinpanion pape buihlmg he .,ugge ted that lightning comluetors he in-talled "at each end" of the buihling.
,\.s the knowledge of the value of the lightning conductor <preatl across the p6 work! but before modern specifications for the lightning protection of st ructures [,,d had been drawn up, the construction. of protective ysterns deviated from 4
'8h, Franklin's original -nauestions in several respects. I.et us first consider the WA ,.1.:s
" lightning rod" or finial. Franklin had 3nege-ted to give this the form of a j; %*Q
" brae wire the size of a coinmon ' knitting needle sharpened to a line point." g \ ,p Nhfd s
($ h
' 0 )q ' i 1 /
r
,\
\
j f'
,,/ \ ' (
y<,(u;{/f V ,
' ' i gjN ,
l t ^ i ? i,
' U l I a w.,
I J S - i W q(/ m J > i\ , 6 l -. }f r I
,[r /
r LJ " V 26. I'm. 7. .\ir teoninals f aiter l'. S. Cutle) (31). lie was clearly led to thi4 suggestion by his obcervation that a pointed eartheti condoch.r i< more effective in di.-chm.:ing a charged body t han a rownled or blunt conductor. We now know t hat, -o far a , huhtning is concerned, this argument is invalid, but it led to the wide-pre.nl adoption of many picturer pte e but splite u, ele.-s de-ig's in whh h a .-harply poitiled rod was -itrroumled by a muhiplicity of spikes giving it occasionally the outline of an angry porcupine. Soine bhune for the pre.-ervation of these features mu-t he placed on the repm t F n.. s U6t nii< ;,rnt-J..-i wrm nim imm vo. x , o,,,, m
.IGS
**+o-e no , , , . . . , , ,
,y..e.mpaa.ne em e e. eengpo gg,+ e===. * **e.Spe.8,
, 90004058
y
.... z........-. .u_..~;.s.,,,.,_,,._, . _ , , , , , _ , , , , , , . , , _ _ , _ , , , _
l The I.iyLining t',mda< for proiinent by t he 1.ninion I.ight ning !!o<l ('onference in which, as inentionni , c:n-i:er, a lightning roil wa still ernlitnl with iln fonction of ines enting the iireni t ence of a hglit ning disch.o ge. t he t he ot her luunl, lu m hele it scetus lias il is r >n.-t ruction licen pre ervnt hinger t han in t he (lniini St:nes (31), as inm 'i by t he ritapes of finials reconunendol as late as 10'>.! (-ce Fig. 7) . .\l. t eriatively, a inoltiplicity of slun t linial- -uperiniposnl on insiivental roof coinluetors for buihlings of laine groinul surface is still reconunciulnl in the i 1.iic t nlition of another .\nierican Unile (35) (3ee Fig. 8).
.\s .itzgesteil earlier, a liglitnitig ro,l cannot he relievl ilpon with certaility to
.tt t ract to it eli every liglit ning .-t roke of low -evelity. ( Figure 10 [p. 303] of
! the c..fup:ution paper by llerger rliows a direct lightning stroke into one of the f0 A4i
.> Q$.*'h4.,,7 v"?%
Spr Ya p3ithg,4 1[e 4-
'l -
,d m
e 4 . 'N 4 # .Y l' II A
- Qs g ,# *N t.
p (.9M . aw // % Js
, , . . p -4 ly ;y L e
' gl
- f
.h g./ M 'g*" -
/g .J ' q
; (
g'%}( .(#g ,h# , #. 6 (* s
. A Apy it x M1#' /
'3 8/7
/c] .
xN x v y"x }/**a i A~$t(/ # i
#f
(.O Ridge cond,;c tor
+C' ,-
/ M / I
./ ,/
s/, - / -
/ \
+.c.
/
.to l'ii.. S. l.i.,htning protection oil.trge roof. f a) af ter l'. S. l'enle 4 35). gla .if ter liroi h ('ude 38).
'! W')
WI. .wi. L n. June thir
~~ *^~'- - - _ , . . . . - . . , . . . _ . . _ . . , , , . _ , . , . . . . , , , , , , , , ,
90004059 l l l
. - . ~ - ..... - _..
IL II. tl No such a ca e t he roof ce inll masts installed on .\ fount S (n Salvatote, aint rimilar "failurei" of a tall wires suspemh d froni li_,hining cotuluetor invariably to attract to itself every lighting stroke have as to exclude t he piis-h"en repiitled iront the l'.ilTel Tower aini lhe lhopire State Ibiilding ) Innidio~ This ~-chenn Atletition has alre:uly been drawn to the possibih.ty of, 3ay, a sharp cot ner . electrical t r:urmia . of a roof 3truelure eatining a point d.reharge:unlg.iving rre to au n}m ard rtreatner w hich then att ract a light ning leader otheke to t he st ructme,19 pan.-mg t he Il I' intere-ting , lightning rod. To avoid t hi ri3k mont inoder n th rommrmlotinaor sinte that the """CC" fully by Spor roofs of buildings he fitted with a ritige conductor running dong the uluile 'luhe duferent n a30s. length of a gabled roof (Franklin'3 " middling uiic") or alonu the parapets or 'luently they Imve !! edges of flat roofs For roofs of large dimen< ion <, the3e coiuloct or . had dong the estro i m y year peritneter are supplenu'uted by :alditional horixontal tapes providing a ine3h of ""P""'ble t o proviih pre,cribed dimensions. Tvpical reconunendations for the width of such a mesh
- ba-ed on nonnal rule are listed in Table IL in-pected in later ye.
, N "r P thus sugge-ted T.e 11. Donmiunv of /.'oof Com/nefor .1/nh lightnin;; protective ;
i f scheme consists of a e"l Maximum reconunemled Ji3t.mee between otspended frotn two t
! mmhn tors fin nieters) uni extending in hot !
i i irdinary sullicient lengths at . Count ry Pef. atru.tures Ihnger otrue'ure" em't hi% l Doern Continett. Saiteerkml (30) 1.5 i37: 20 to ()n & a lbhinin, tkrmane ts 4 0 feet) L*, t o ". o ! .G to 10 feet) an or. ting
!!ritain ' 9 81 it is the ftmetion to Jegree of ri,k )
i35) ' 50 (with mldition.J finial 3) cart h electim' U. S. j for a hni;< ___~_.._.__.._.________.____ ennduc t ore Following the foregoing con-iderations to their logica: conclusions, -ome down condnetors i . modern Codes have d!< pen 3ed altogether with the provision of vertical fini:da tlos si unacceptabh-for roof st ructures Thi., dewlopment i3 based on the arumnent that muler the ,ggg
' inlluence ni the strong elec tric tields di-cit-ed callier, the shape si t he c"n-y,> Ilowever' f.or ma ihtetor dengued to inteleept a lightning di3charce i4 immaterial, i ciine!'.sion
'- which is -a,. ported by the breakdown vo!! ages of long 3 parks undei ana-:ronted tall -t ructu such Unpul-e vnhages. Thu<, in the latest ilritish Code (38), the prm :.-ion of an e mductors are rerpf array of verticallightning rod 3 i.. recommended only for -mall danger 3t ruct ures crete hmidings, the while even f or tall chimneys the provi.-ion of a ring conductor in-tailed a!ong sy-t em provided .t i g g the edge o[ lhe top o[ the structure 5- deemed to provide adquate protection, Neither in the Swiu (36) nor in the German 130 3pechications are vertical pen o with -eparate tildals luenliinteil at all. Tln' only faulu tif einnst ructilen uh$ch runs etiniller to h..nd the upper end-t hin development is the so-called mlio-ortire lig!dnieg com/vch for w hich it is have i.t connecte I to claitned that a
- ingle vertical cotuhtetor enn protect a huilding, of largo 3utface I
area. Furthi e teference to this device is made later. Inei.n ci, th.s -c! i w hich goi- hevond for c%plowive stores or ol!nT similar 3ma}} ht rnetWem it W ay be im ptrative
. hones Clerk .\f aw i to prevent a lightning current contacting any part of the mtrm ture surface, in i
i
)
i jseur n.el ..l Th6 i' f.Ua IM) 11a m dt f vnI fil, Sp 6, j u n g l'M
' ' * " ~ " * '*8 * -** ==e.e . . em g
*+9
B'*W**'"*#""'"****'I
'**8WSeg'*.
- .g."
- une
- 90004060
. _ _ _ . , . . . - ._m._ , . _ . .. . _ . _ , _ . . . , _ . . ,
a a 6
)
,1 The f.r'y!,Inhnt t%dnt< r such a ca-c tlu roof con.luetor systein i, best icpl.u ed by a 3 rtem of catenary ,'
wire -ngended froin t all tou n> :n ranted :u ontnl t he nirnet ule anil 30 de igned a in esclude t he po-sibility of -ide lla-hing from t he proicetive <yricin to t he . i buihling. This -cheine is based on s he saine principle :c, t he pintertion of an eleeliic:.1 t ran-ini--ion line by an os ci he.ul gionsul-wire.
- i f
'.t is ittleresling to note in passing that the saline prilleiple ha, heen adapled ne .cce-nfully by .%por 139) for t he protection of rinall farinhouse, in 1%laml for ipUt e dilTm ent rea,ons. .\s a rule the,c buildings are made of wooil and fre- !
ipa nt ly t hey have t hat ched roofs. .\ large munber of the-e farmsteads are , de troyed every year b3 lirr started by lightning amt it prov"d economically imporible to provi le all these struct ures with cliective lightning piotection ba-cd on nonnal rules ami diinen-ions and to have t he-e in-tallations reauhuly in pened in later y ears to en-ure that they remaine 1 in a -ati-factory tate. ,
.hpor thus suggested a scheme which enabled the farmer to install his own lightning protective ,yntem atal to do so at a minimal cost. E3,entially his ,
eheme consists of a galvanizeit iron wire of 10 nune (0.010 -<t. ind cro% section
- 14pemled from t wo u ooden upricht, in,talled at the two ends of Ihe roof ridge u.d exte:nlinx in both direction, mala a -loping angle dowe. to aroinnt level,
, n!ic:ent lengths at either eint bem: hmh d in the gnnnut t > provhle etfective [
eai t h mg. l i Don n Conductors On e a lithining tt oke has been intercepted by the roof ci inluetor syeteln, it ,, t he inaction of the dou n com! actors to t ransfer the light ning current to the , eart h "h et rodes. For < mall buihline a single dow n conductor s ade<piate, but i
.or a buibline of large armnni -urface and for a tall structure 3everal down {
"o nhciors are res piire.1 if ide :la-l.in to internal metal i- to be avoided. .duch '
down emnluctm < may be estal'ed on the outer -urface of the buibling or, where t h;, is unacceptabh for asthe tic rea-ons along internal walle or tertain 3ervice ! l.aiH. Ib.n cver. for many modern indn-t rial or llat bmklings as n eil as for many ' t -d l .t rnet urm. ruch .., wat er t..wers, cooling towers etc., no reparate down e celuctitr4 at e re<ptiled. Thu , f<ir -leel framerl -t ruct ure* and reinfierced con-
- i ro'e buibline, t he internal na tal can be utilize I as part of th protective
+ tem nrovided it circ 4 a direct metallic connection between it uppermost vel lower enil,. .dimilarly. hmbling, u n h "ontinuous em tain namn can dis- i
- p. n-e with +parate down conductor- .\ll that is nece--ary in ench cases is to i bond 'he upper ends of t he metal (nunework to a roof conductor -3, tem and to l hwe it connecteil to m etheirnt earthing :ystem. :
in cliert, thi- -cheno lead, to t he oldy ilopoi tant light nin Un icetive lom ,
'hirh u e bojon.1 Ft anklm'- innuinal idea- Thi, cheine lir t -ie:gested by ,
.h nas l'lerk .\inu cil 4 40) and now applied to many small 3truct ure of ex-I h.i w% r., J ne eu f
- - . . ~ . ~ . . . , . _ . _ . . _ , . _ , . , _ ._
l 90004061 i
_ _m
. m .. - . ~ .4._ .. . _ . . . _ . . _ . .
R. H. (l alate In me explosive liazanl, ut dixes t he conception of the . o-c:dled //,n oday roue to Fir-Ily, ap;o t fium t he im avoid the ocem rence of any potential difference within tim pnitiriol buihling conduct or, lim olonm volt:
.so as to make tin buihling, in Manvell's own n onl , "a clo-ed con.locting ride thishing amt secondly ve+el."7 It is t he development of 3nch a potential dilfen nee bet ween tlm Ught. regt function of t he muf ning comluetor spiem :nnt intern:d meial, carthed or otherui-c, w hich has numy fatal accident, to lo given ri3e to the great majority of aHeged failures of tim protective -y stem aml w hich brought Franklin's nystem into early unjustified di tepute, linried plates which h:
present time have the dr: The ri k of side flashing can he overcome by emeient multiple honding to wideh is liable to drying i the down conductors of anv extended metal fixtures or services in or on the Imtueen plate and nurnu building. With increasing building height. such bonding mu-t he done both at w hich compre-, t he soil a. the highest and lowest points of any extended metal component.- This al<o be u-ed because of du roc
,tre-es the need for cateful record keeping and periodic checking to en.-ure ate ellective, partienholy that any aheration , to ervice pi i* and other metal fixt ures have not interfeied iheir initial surge impedm adversely with the cHicacy of an originaHy sound arrangement.
lionding of the earth Much uninformed critici-m has been voiced again-t the use of a right-angle metal services such as e: bend in a dow n conductor or its connection with a roof conductor. The hending method not only as.<i.-ts force, ari-ing at such a point an u uaHy too low to a6ect a cominctor of u nal prote"tive s.v-tem but it crou section .uul ri3ks of shle da hing.are ordigible unic, the runductor is bent groumi. Such bonding m back o as to form a long hut narrmV re-ent rant loop. tion,but meansare aYd L 1:arthing System The Protective Han Praak!in we wcH aware of the i:aportance of a gwl carth connection for a Information on th light ning ro l. Already in Poor !!ichant's Ahnanac does he recononend 04) light ning cominctar st ruct ure, is ( " ! that the light ning conductor he "of -uch a length that at one emi heing three or l four feet in the moi-t gromul." In later publications and letters he revert., re- ! peatedly to this subject. Thus, in a letter dated Fobruary '/Oth lW, in which j he comment., on an account of the -occes ful di-charge of a lig'a ning tiarh to ._ - . - - - - i the hou. e of a Mr. ht in Philadelphia, he remark., HI): I Ien' N llk}e ( $k!N' anCe, r . I'{f ( (e [I !! ll u El*$ Ni'{\ kI) ( l-c blMllkr[ itself from the foot of the roit over the wet pavement, which seem4, I think, _-.--- - to indicate, that the earth under the pavement wa- very dry, an I that the ag n.d -houhl have been sunk deeper, till it came to earth moister, and t herefore itrmin apter to receive and di-3ipate the electrie fluid." U.dand ,
- oath Africa l Inade.ptate earthing of liehtning conductors led to many faih.res in earlv - - --
l in-tallations with the re.< ult that this matter was given -necial attention bv the ' Il " '" * "I P'"t "I br u keb h..s e lin n denvra 1rench Academy of :. .ciences whorn wrien of liccomtnetulations have 'been mentioned before. A low cart hing re istance nny be reipiireil for two rea,ons. aul to nia in national nm in ubch tho stati i for the som rea <on per" n in :oi .il'-el var or avoipl ine su e inumem from elect rie , , _,g, [g hoi k. Mnu eth ov identallt, w as d-o t he hr-t per-on to dre n at tention to t he f.o t t hat a tight nsg condui tor w ould at o.u t more lla-l.ca th in u nohl hn e o. . ormd to tls -an.e pot Ph4"'I I # "Y " h:nl it not been titted. vn pn e -mee the aullo
'pI *= juiH tt.d 4H II% l (.olk hu l%DOdc \ eil. l$I, b 4, )MN NAI
..,~m.,._-__ _e-==v-*-'-*-# - " " " ' ' " " * " " " ~ - - - - - -+ --
"**i- .".eug. _
8*- 90004062
3 1 _ . __m_ . _ _ _ . ._ _ _ . . ____m _ _ _
._m_ __ .. . _ _ ___ . . ._ q I
TI,e LigMning C mductor no-called Faraday cage to
- Firstly, apart froin the huluctive voltage drop which arises across a long down n the protected building conductor, the olonic voltage <lrop in the carth electrode detennines the risk of b, "a closed conducting '
side flashing and secondly the potential drop across the ground surface is a di-rrence between the light- rect function of the carthing revist:Utce. Tho resulting voltages have led to or otherwise, which has nmny fatal accidents to huinan beings aint to quadrupeds. he protective system and l ' disieput e. Duried plates which have been widely used for earthing purposes up to the present time have the drawback that they have to be placed in made-up soil ient multiple bonding to i which is liable to drying out so that, in the course of time, the area of contact or services in or on the - between plate and surrounding soil may be materially reduced. Driven rods
'g must he done both at which compress the soil are therefore nmeh to be preferred. When there cannot I components. This also be used becau<e of the rocky nature of the ground, horizontal buried conductors odic checking to ensure are efTective, particularly if they are arranged in star formation so as to reduce tures have not interfered their initial surge impedance.
gennen t.
. Bonding of the earthing system of the lightning conductor to all buried
; the u<e of a right-angle metal services such as electric cables, gas or water pipc3 is imperative. This conductor. The hending method not only assi ts in securine a hav overall earthing resi.-tance for the
'ect a conductor of usual protective syntem but it also prevents the risk of a long are discharge in the less the cotiductor is hent ground. Such bonding may tend to interfere with systems of cathodic protec-
. tion, but means are available to overcome these dilliculties.
1 I The Protectiec Range 1 21 carth conDection for a Information on the protective range or protective angle attributed to a l n he reconunend (34) lightning conductor in several modern national Codes for the protection of ' t one end being three or 4 structures is collected together in Table III. Considerable difierences are seen
- nd letters he reverts re-Tunx lI1. Protective Ranga (or .4 nuled ary 20th 1762, m. win.c h Adopted in National Itecommendations e of a lightning flash to 1): l l Ordinary structures I) anger structures l 6 l ng was seen to difTu e country iter. It/It a It/II a l
, which seems, I think, _ ._ _ __ _ _ _ _ rery drv, and that the l
'. IL S. (31) ti 1:1 (45') I mobter, and therclose Ikitain '30 fl'll C2 ' nW I) 3u*
. Poland (42) 1.5:1 4 South Africa (43) (1:1) 4Y i many failures in early '
special at tention hv the It = nulius of pmtecnxi circular base. Il = height of lidining roniluctor. I'igun-+ in brm kets have been deriveil for purpo-es of comparison. 6mendat ions have been 1piired for t wo rea-ons. Mill N & i di d n mm Wie TW Nitia Cd is N o$ de
,e are inunune from etc.tric ment in which the statistical aspect of the attractive range of a lightning con-I uttenti..n to the fart that a ductor .is stressed and m. wh.ich .it is . stated that a single lightning conductor 1 occurred to thn >;une spot ,
cannot he expected to provide complete protection. This fact need not cause l t surpri c since the author must admit at having exerted a certain infhience in I Journal n( The I teLhn linniute Vol. M3, No 6 junc 19d 'I I ' ,3 l l
- i 90004043 i ,
1
.. . . . _ . . , . - . . . r .. . e ,~. ~ 3. u,.-. ~ . ~ . . - .
r' ,
R.11. livMc , -
the deliberations tif the drafting conunillee. ( ht Ihe other luuni, it is noleworthy then recorded the ]
that the Swicribing to any t hese investigations liz,hining protective system a limited range of clicetiveness. Completely different gradients (di-tant claitas are nuule for prolcelive syntetus eniploying radio-active materials in. point disch;u go corporated at the tip of a conventional vertical lightning rod. The sugge tion to currents of the om.
use radio-active materials for dius aspects of lightning protcetion is by no moven closer and ,
; means novel, but claims have been nuule in tecent years to the effect that a Ifillehrmal conclude.
j -iegle rod of normal height, if litted with a r:ulio-active tip, is capable of pro.
; tecting an entire buihling of large horia,ntal dimensions.
,I,hus, i.f it u. adn i In order to te.<t the eHicacy of a radio. active lightning conductor, Tintier. lightning conductor Ilillebrand (44) crected four of these devices in accordance with the supplier's conductor must be t instructions ami, nt distances of between 3l atal 53.1 meters, he erected a con. tional conductor of ventional lightning rod of the same height as the r:ulio. active conductor. Ile ;
c) . Tlu only nation-b) 2 10~6 p - --- - -
- - r j j- p r active lightning co l provision of radh
- c/,
10,5 ---- - worthy c rfect, Th.- rm J 1 i t A p !
- 5-
, scientific investigat:
{ --+- "%r e )g, , natural ionization o { co active material." Th
=/ e .
i i i 2
"[f/jfg-- ,
f l f w 2h- are known by which conductor, 10.$ -- j ' 5 ,L5 - -- L_ b - . L , i-
.. . l,,.
e Conclusions a 2 p4 -_ . - ,
; -[ J ,
a T1.c mode of - iished. Its - ' ,
/ 1 ! . i ': e 4< ,
i ;g 4 , contact a; 10 7
- 7 p
p, current hau
/[
- - i 5 [~ ,[b, l'_. j L_.4- - '
[ . _., nietal can he over . of aido ga hing a a l t f)c._7 -l4 , gle,b l l
, ! , : protective erstem.1
,y , ; reei3tance i4 the red-l 10~ L8 ; -
'MT-- . C., th point of arthin.
l3 I I , i . e"ential hetween t: 5 -a w - L'a L.r._ g' e0 ! hurica metai pipes. I m, <v , o ! I 0 0 i 2 , 4~ - r-t ! Apart from chur I , j , I i roof conductors whi. 10.g -
""I #""'"" I n q I
0,1 0,2 0,5 1 2 kV/m 10 0,1 0,2 0,5 1 2 kV/m 10 conductor.1;or hm'
+ E -+ - E ----+ - tinn. cmuin wg Fu t 'l Variation of einioinn current aith elettric gr.ntient : tai po, dive gradient ; o negative 'tructural nu ial is e.
gradien t ; l ,2,3,1-radio-:p live li,:htning condatner, .">-normal lightning a dortnr. conaluctor 4y3 tem. Juarnal of The ihnl, tin lediute vnt ."u, h 4, June 19t.1 5 '6 4 6 90004064
.... ~..
I I t
?
TI.c 1.lyhining Gmdudar . [ r (lu n n corded the point di.,rharge current. produced by t he tuo cinnluctors j innler tinnaler.-tortu condition, in Switzerland aint in Swnlen. The re-nlu of t hese investigations arc ounnoarized in l'ig. 9. The curver s how that, for small j gradien t s blisl a n t, timnder torius), the normal roil produces no mea in able l pomt discharge current, whereas the t adio-active light ning cotuluetor di< chap;es current s of the order of a fraction of a niicroampere. llowevei, as a t hunder.s torm , n e rv es closer and the grailient incicases, t he curve, begin to merge. .Tlilller-i llillebratul conclude.s that this result is in full accord with theoretical prediction. Th u 4, if it is allmitted that the current discharged into the atmo-phere hv a hghi ni'ez conductor i, indicative of its attractive action, a radio-active light ning [ coinluetor mu,t he taken to provide the same degree of protection an a conven-tional cominctor of the same hei::,ht abete ground. t 4 The only national Code which appears to nake <pecific reference to radh>- l aHive lightning conductors is that i sued in Germany (37). It 3tates "The l providon of radio-active material on lightning in cannot produce a note- l v orthy cliect. The adihtional ionizati >n created by them lies, accouling to [
-rientine inve-ti:.ations, several orders o.f magnitude bebnv that canned by natural ionization of a lightning rod in a thunderstorm tie!d wita no radio-
"tive maierial." The British Code (38) ou iely indicates that no artificial means f
- ue known by which to increa<e the range of attraction afforded by a lightning i">!nldellir.
t t
! 'o n clu sion s b The mode of action of a lightnite.r conductor is now reasonably well e3 tab- l.
li-hed. It4 mie purpo.-e is to intercept a lightning discharge before thi.4 can i' contact any point of the buihling to be protected and to discharge the 'ightning current harndeely to grouml. The risk of side flashing to internal or external linend can le overcoine by bonding and, i[ thi3 it e@Ciently carried out, the risk
.if ei lo dehieg i, un:,tfected by the magnitude of the earthing reni3tance of the ,
puiteetive sy-tan. Under su h condi< ions the main a<lvantage of a low earthing i ic-istance i4 the reduction of the potential drop across the gn,und surface about { the point of oaithing. In order to avoid side Ha-hing in the gmund, bonding is l e"ent;al between the lightning protective carth electrode and any adjacent t
!aur!ctl niet al pipe:i or cal)leS. f l
.\pm irom church steeple <, vertical finials can he :eplaced by horixontal l rooi co:nluetors which rhookl be so airangni as to cove r all -harp edges of the .
linif t ruct ure [nlernally alf angN1 dt >Wu colnhiets tr< are .14 et[cetive as eNtcrnal f c nduct ors l'or bniklings with steel frann- continuous reintorcement or con- (
! !nuot t 4 curt ain %ahing, dou n CondUrlorF enn he di.'peDSed uith, providMlthe I
-i t uct or:d metal is etYect ive!y eart hed an 1, u lo ie advieuble connecteil t o a roof t eionluct..r 4v3 tem.
~ l k
- I ')
V.il. :U, W. c., j mic i .9 t. 8 y- - - - - . . , . . - , . . . , . - ~ - .~.-..-.....--we-** 90004065
R. H. Gul,le Tin. 3 pace pmtected by a lightning conihtetor i< .still sul>jeel to fin ther in. do) T. Ilonath, "rine n ve-ligation alt hough t hes e are ni rong !ca-ons to believe that the di 'h N""'" N '** \ l w h. ic h a lightning conchictor m capabin 01 attracting a lightning schaine u aih.
-tance overG L ) .1F. NS,Eh,wGib ah, "Ih re, le
*,c., p tii nialistical yttanlity relainl to t he intetisily of t he lightning st toke. r2O IL l bvis, '14tniin- i Q D 8vc lbf. nih I.ctler.\ !
In concht3 ion, the author wishes otice Inore lei pay hornage to the genius of ,34) g , g ,p,,,n, o g e p, g g , l llenjainin l'raliklin w hose li"hinitig colnhlctor systein lef[uiteil only tuitior p. 7.n,17;s.
- ulditions or mmlification.s since it- first enunciation in 1753. Truly, he was over- G5) "in iruriiom ur le,i inodnt in writin;; (45) in 17C "Indeed, ist the construction of an instrinnent #20 IL Andumon, "Lighti i27) W. II. I'recre, "t in d so new, and of. win.c h we could Imve so little experience, it .is rather lucky that o . U. .,.,1850.
we shouhl :tt first be .m near the truth as we seetn to he, and conunit m few Gm . W:.her," Von wo error.5." l'h yd, Vol.18, p.10 Q9J H. Wd3on, "On the Hil>liopuphy ')* V' " ' Y"*k' " UI'I" liiil 16ok ('o., Inc., I (1) K. L'earson, "The Ur:anraar of Science," London, Adam & Charin Black,1911. (311 U. S.1)epart na nt of i (2) IL F. 8thontand, 'The Work of Benjamin Franklin on Thunderstorns and the Develop- Pr:2,1937,19 65,195; ment of the Lightning Rod," Jour. Frank. Inst, Vol. 253, p. 375,19; L 02) P. ( ; Provoost, "Th Q) L 11. Cohen, " Prejudice against the Introduction of Lightning Rod 4," Jonr. frnnk. No. 3!4, Appemlix 11 Inst., Vol. 253, p. 393,1952. Q3) C. F. Wngner, G D (4) L H. Cohen, ' Benjamin Franklin's Experiments " Harvard University Pres, Cambrike. Trans. ,1 J.E E., Vol . Maw, 1911, letter to it. D'Aihbard, of June 29th 1755. <34) .ee Rei.:4 6 p. 129. is) i L J. Sy mon., (editor;, "th;htning Hun Conferente," E.1 F. N. Spon, tendan and New Osi Natic.nal Fire Protect ' h rk,IM 1 06) .bociation Sume de 4 #6) () liver J. Lodge, " Lightning Conducters and I.ightning (;uards," London, Whittaker & tion contre la (nudre,' ' Co , IN)2. (37) Auvehua fur Blit za (7) J. Alan Chahner.,, " Atmo-pheric E crtricity," Pergamun Prc:.,, London, Paris, New York, Ucriin,1963. 1957 '3M Driti<h r t nd:> rd, Ian i8) B. F. J. 8thonland, "The Lightning Di*harga," llamlhook of Phpics, Voh NNII, don, 1965. Berlin, Nrw Yor k, springer,1956. 09) 8. Spor, "Paratmo - I ! 4) C. E. R. Brua and H. IL Goble. "The Lightn;ng Dischargo," J.l.E.E., Wh M, Pt. II, p. 263,1959.
- p. h7.1941. C. M. IL Druce, "The liotiation of Long Eien trical Ibchar;;m," Proc. #vy. (40) J. Clerk U .
&c., Vol. I S.1, p. 228, lu t I. port, Li (10) IL IL tiable, 't he Freipiera y of ()crurrein e and the Dist ri:iulion of Li.;htniag Fla-hes :41) See net .4 , p to Tran mi% ion Lice 4 " Tres. .t ./ K E., Vol. 64, p. 402,19 65. ,42) Wy.lawnictua No.
(11) H. II. Gohle, "The Attractive Erie < t of a LL:htning Canductor," JJ E.E., Wh 9, p. 212. Wa rsa w, 195;).
)
- 1963. 43i South Afrirsn Burent l r12: A. \f. Thnma.<. "The SaitchiegAurge etreturth of In3ntating Airngements for A ems Apin t [ :ght nir".:," t
; t iperating at Vohn.cs \l'ove 100 kr,
- E R..l . Report No. >0. I cath rhevi, P 66. (44) D. .\lnller Itil:gbranei i t iM T. E. AMibone, ".\Ierh:u.i m of the Elet t rical lireakJuwn ni I ar.;e ( Lps in Air," ('/GRE durch Haundm uneen Heport No.128, Appemhx I,196 L (45; see Rei. (4), p. 373.
(14) L 8. 8tekolnikov and A. V. dhkib v, "Inve.-tigation of the h hani3m of the Negative Spar k," Xad NMR. Wh 151, p 1l85,1963. (15i IL C. Ih.;hes and W. J. Hohert, "The Etica of Wave Piont. :)nration on the impulan flehm er of Air Gap , ' Pv. l.E E., Wh 112, p. los 1965. (10) T. Udo. "8witihing Surce and Impul3e 8p irkover Chano teri tirs of Long iiap Sp.a me md Long in.,uhtor 81rin;4 * /E.EE . pn per Na T P-6.b10 l. (17) E W. Boehne and i;. ('irrara, *1)n the E.lLY. Nitching Surge In nlation oirength of Lme and 8t ation Insul .tmg Air Spares," ('J.(,'RE., p iper L 115, IU6 L (181 l.. Pari <, "In'hwnee of Air ( bp ( br tetere.tii n on I.h:e to (;round A dching 8 urge Strength," .1 J K EE Pm r 31, pp nG 121,1966 (19) K. herger and E. Voakovr, "l'hotogratiwhe Bliunnteroa hine,:en h r J h e 1"55-I"65 auf dem Mont e 8a n s .dvatore, ' N.J/. .l N L, o. , p. o90 1966 b Journal of Ti e I rwkhn ins 6tme v .i. m L. 4, June sw
. - . _ _ . . _ _ _ _ - . . _ _., _ __ _ _ _ _ . ~ ._ ..
90004066
/
, ... _. . - ~ . . ~ u .,_. a , . ~ . . _ . ~ - . - - - . . - - - - -
Tlee f.ig!dniny Coinludor QO, T. llorvath, "Eh.e neen \letluule inr 1:rinilllung <ler Wahr chiinli.hkeit von Rh'zche a bl. w n," Kh t.h se, Vol.17, p. 216.19n.t.
# 21 i 1/. 8elorali, "l'erci l.nnng iler 8ebut /w ir kung; von Ithi.mbiciter n on.1 Thrmen," /!aII.
1 X E., \ of. 36, p. 67X,1965.
,2;i it.1).n i.<, "1. cht iong l'h-hover on t he lirit ir.h (lri.1," /%e. /.E E., Vol. I lo, p.969, 196.l.
4 0.0 See Itef. (4), i.et ter X X I V, l'ari>, p. :M, sept.1767 r 2 D 11. Wil-on, "ltepoi t on 1.ighining .\renlent at i'urileet." l'h rt. Troin. I,b t . k*., Vol. 6M,
- p. 2:19,177%.
i23) " Inst ruction.4.-ur les l'aratonnerres," Ae;elentie Ito3 ale sle3 science.s l'a ri. ,182:4. r:6 It Ander on, "I.ight ning ('ondni tor *," I:. & l'. N. 8pon, l.ond ni. New York,1879. 2h W.11. I'recre, "On t he space l'n,te.1r.1 by a lagh'inne (' n.due sor," Ptol.1/oy., Vol.10,
- p. 12 7, I M
'28) 11. Waltr r, " Von wo ali steuert iler Illite auf .-cine Ein3rblagarlle zu?", %. fur Te clin. .
/'Apik, Vol.1s, p.105,19:17. I (291 U. Wit-on, "On t he Termin.ition of Coin hn tord," l'lol. Tro,or., Vol. 6x, p. 999, 177't.
r30, l'. W. thi, " t helectric l'henoinena in ll;gh.Voltaw Engineering," New York, .Wt: raw. ( Ilill Book Co., Inc.,19?). f
- 31) l.~. S. I)cpart tnent of Conuncree, " Code for l'rntection Again.st !.ight ning, ' Weh., IL C., !
19:12,19:17,1911,197>2. e i 3 D P. ( 1. l'ro voo- t , 'The shic! ding 1::Tei t of Overheml Eaith Win , ' F / G E E., paper ! No. :ll4, Appenih s !!,1960. !
'3h t i. F. Wuner. ii.1). .\lrCann, t ; f.. Al.n f.ane, "8bichling of Tr an-nti,.-ion Iine*." ,
Trnos. .I l.ILE , Vol. 60, p. ;;t.l.1911. , 31; 8"e Itef. f 41, p LJ9. }'
- 331 Natinnal l' ire 1%terhon b ori sti in, "I.ightning ( ble 1967,," Co-ton,1lu.s
'361 b. o< iation suive des 1;lectrb icn- "Iteromrn un tation, pont h 3 in-i d'ations de protec.
ima contre la ; ,udre,' taich, lum i3") Aue,chu% f n. Bht zal,leiterbau eX. ' AllD), Allgemeine lih t /.,thut elle-t irninunw n," lier!hi,146:1.
- 33) P,r:ti.,h 8tandants in.-titution, 'The Protes tion of stro,(nres Against L:ghtning," 1.ou- l
< hin. I M5. ;
3 A S. &por, " Par tonner res ruraux de t) pe h'ger," l.', ar Widrofe de l'K/,,(ned/, Vol. 68,
- p. Jil.19.*;9. '
;40) J. (1erk .\laxwell, "On the Protection of noildine fnun Lightning," liniuh Ann.1%
part, London,1477. I el) 84e Ilef 41, p. :17 2.
- 42) W3 dau ru t u a Normalizaeyjne, M h hnina Pavlowli o l Wy!.ulowan At mo-feryczu3 < h,"
Waua ,1959
.4h pouth Afrn an Dureau of 8t.o.<hnis, "Onle of Practice for the Protection of Ru! dings .
Again-t I.icht nm.:, ' l'retoria,1959
' 44i It M8er-!!i'.lchrand. "Bectn!Wong der lWt., bahn . lurch radio tLtive 8,rahle i unil '
dun h it nanl ulanen." Ele'.h oto unVbr E v wAn.o. A. Vol M p.152,19ii2.
! 45 i see Uci. ,4), p. :17:1.
l d
- m. ,. m. - m.
- e. e - w w ,,, w . - -. __ -- -% .. ..e --
90004067
M. f2 L F, 4 4 The Relation Between Stroke Current
- 4. Each wave of charge q,, in coulombs per centimeter <cm>. s s >> c1atea ->th i. a
{. wave of current s,, in which g. and the -Velocit of the Return Stroke "
- 5. The are plasma behind tha heads of the C-waves has a constant radius a in cent - f meters; its resistivity is zero bi C. F. WA 3NER e. All of the current sows on the surhee pl N#E of the plasma; aJl of the charge IM:s on f, its surface. i t
W mary : The velocity of the return 7. In Fig. 21 of reference 8 it is shown ance of the path of a discharge varies , s- ie is an important element in estimating t. versely as the charge that passes that that to bring a 6 foot spark discharge to j .,o O de surge impedance of the return stroke, high conductivity for curren ts between q. [ (! the potential of the downward leader, point m the discharge path. It is the 1,000 and 2,000 amp requires an energy ,
; c .-
and i3) the length of the last striking dis- purpose of this paper to re examine the that is proportional to the final current of , t2::m The energy required to establish an the discharge. This energy is approxi. },/ relation between the stroke current and a plasma can be determined from labora- "" " trr tests. By equating this quantity to the velocity of the return stroke in terms centimeter per ampere. \ en6 cation of this ,I " the energv required to retard the velocity m te susceptible to physical interpreta- value is given in the Appendix. tion and more readily determinable from af a traveling wave, the consequent veloc. The properties of a system of waves J : )[. irv 4 the return stroke can be evaluatedin te _s of the stroke current. laboratory tests. such as the one shown in Fig. 3(A) have been discussed in connection with Fig.12
]P
- Simplified Analysis of reference 9, in which it is shown that ' h;.I g N.
i presented an inter- to establish, as has been premised, rec-u To develop the physical concepts, a 3 6.', f I,,-s,d. ting .relation Lundholm between. the ettrrent in, - tangular waves of charge and current number of simatifvi=E assumntions
- will I,.; L which travel with a constant velocity of a:., the velocity of, the head of the return 1-be made. Later these will be modified or sd e- Rusck3 later reviewed Lund- vc, series-forc.ing voltages, as shown .m ;
n1; Jf eliminated by further discussion. These . . o he-m s derivation utd, af ter modifying Fig. 3(B), must bc inserted on both sides ,j l, assumptions are as follows: h one of the constants, arrived at the fol. O and 0,. In addition, a retarding m-ler.n; result . L m nturn suoie can be viewed as a voltage must be inserted that is symmetri- , W j. wave of positive charge (and current ) cal with respect to the heads of the cur- M-4 I 5 0 X 105 wh se ead uavels upward, neutrahng. (g) as it progresses, the negative charge ! aid rent waves. The paths of the waves are ,
- t. : /\1.,- l down by the dowrward leader. This wave separated a distance D; it will be shown M;i 1
(L
.. . moves upward rrom a plane surface, which subsequently that this system of waves is i.n 2nich t is the velocity of the head of tae return stroke expressed in terms of in this paper will be assumed to be per-fectly conductin g. For the initial as-approximately equivalent to the case in )},1, 1 sumptions, however, the system of simul-which, w hen D is made equal to twice the .
O ths velocity of that of light, and I is the b str le current in amperes (amp). This taneous waves, shown in Fig. 3( A), that distance the bead has traveled from the qIf ,
. . . . moves outward from 0 and '>', will be earth, a wave of charge rises vertically : t rc v.n is show n m. Fig.1. premised. This is done pnmarily because .; i from the earth. Fig. 3(C) shows in more I Fig 2 shows the distribution of occur
- this system, af ter trie heads of the waves detail the total voltage, using 0 and O' I re:x curve of the velocity of the re- have progressed sufEciently, leads to a as ref erence points, that must be inserted Enite solution. All of the essential elements tun stroke determined by Wagner and progressively in series to bring about the y,g
.WCan nt (Fig. n0 of reference 4) as a of the actual case reside in this one also.
All of the waves cf current (and charge) propagation of rectangular current waves. 7.; ig resdt of their analysis of 19 records 2. These voltages are the integrals. from pruented by xhonland, Ma'an. and of Fig. 3' A) are rectangular. these reference. points, of the electric h,p Cccens' Fig 2 also shows a replot of ,U- 'he waves travel with a constant ct a tield,s longitudinalcomponent, at a radius pt } the frequency of occurrence of stroke velocity w, where c is the velocity of light c ents as published by the AIEE and F a numenc. G from the axes of propas;ation. The ig.[
.j,.
Li:htning and Insulator SubcommitteeM d Psk, accepting, a priori, that the veloc-ity of propagation is a function of the o3 ' ,, Q
.h !
{. stre ke current, plotted the mutually con. neced points shown by the dotted line in Fig 1. It can be seen that the rela-tr : between the two curves is remarkably [ o4-
- 0"0 9 "'50" 4'5'*'Ib \ f / '
v' 4k f}f (j g< s d b gh 03" y . g'n 's'M8 7
;h s
;Qk
, f The derivation of the Lundholm expres- *[
us o i h/ U h. g sv.:is somewhat obscure in some respects N P'l we o2- ' /N-FRCW FIELD DATA t as it is based upon an empirical relation
. I y f/
d@i [l yk 6f Toepler? which states that the resist- / L , q6 4 Paper 62 to04, recommended by the ALEE Trans. mm e and Distribut,on Committee and epproved )
; 8, ,
br the AIE E "rechnical Operations Department for , I prewnt.oon at the AIEE Summer General Meet. O so
- loo Iso 200 5
ice (se n ver CoM , June 17-72. 1962 M a n uscn pt
.u t. s e t ed M arc h Ot. 19ti7. made available for va na May 2, 1962.
o STRONE cV##ENT 'N KILoAM'E8ts O('
,I 3
- l. g-u 8 Waowse is mth Wesu n t house Electnc c cp.ratmn. Fast Pit t wurg h, Pa fi S .1. Relations between stroke cunent and velocity of return stroke l ', ! k r
w wen 1%3 f t'agw -Stroke Curren! and the l'elocsty of the Return Stroke 609
'00 ,
surge impedance will be 1/r as great and j 3 i VELO 0'TY I the power input P will be
~, y oc
{ cr. 20 of <ef 41 P=sP, (9s I o The electrostatie energy will be distrib-(3' l uted times as fast as for = 1 and, there-30 y!$ ' fore, steolc
- , a -
- vancur BY 4o I P, = rPu = 5P, (los i 5,v$
\
v5 The magnetic energy per umt length will
- be t2 times as great and will be distributed l [$ ro ,
l v times as f ast. Therefore, l U o os ai c:s ca e e es o4 P.=rP.,= 8 P, (11) v
! o le 2c so ao So 6o 70 Bo 1 *a aca5's The power absorbed by the retardation effect is proportional to the instantaneous Fig. 2. Relation bet =een frequency of occurrence of lightning strokes (AIEE cun ) and niocity of return stroke back emf and the current, but since the current exists only on the lef t-hand side of the retardation voltage, the power absorbed by this effect is 1
forcing voltage in each conductor, af ter develop the energy relations nvolved. i a time slightly in excess of the travel For t = 1.0, the power input ina e $' P, = f I',i, (12) time between the conductors, attains a conductor P is equal to Pis. It can be i constant valae of shown that this power is expended in Remembering that for tdl, i, varies as developing the magnetic energy per unit r, then this quantity is proportional to 1 D l'- 60i, - In - (3) length P.,, which is equal to Li,V2, and
* " s- P, (13i the electrostatic energy associated with Pr = kt'
~
where D is the separation between the the radial field Pu, which is equal to two parallel paths of waves. The retard. Cr2/2. These quantities are equal. Equating the power input to the sum of Therefore, for r = 1, the stored electrostatir. power, the stored
~
ing voltage P, has two components: (a) that occasiened by the charge, which magnetic pow er, and the retardation P, = Pu - P., (7) power: attains a limiting value of and 3 3 p F, = - 60i, - 1r. - (4) Pu = P., = Pc/2 (8) rP, 2,P,- 2,, P,2-r-t rP, (14s r a I which is an identity and veriftes the ulti- l and (b) that occasioned by the current, If r *1, for the same forcing voltage an which attains a limiting value of equal charge will be distributed along mate distribution of the power input. j the conductors, but the current will be Returning to the application of these p relations to the immediate problem,
! r, = 60i, e in - (5) only r times as great as for t= 1. The a
[ The retarding voltages are then the sum of these quantities: (A) 90E4069
's.}_ ' '! _ _.d _ _II. .__ .
i, l', = 60i, fr b in E (6) M No s l 'I o 2 \ r/ a I wo 4 o Let it be assumed that the heads of the o"%
- p o-
', waves have progressed to a point where (B) b wo 2 the forcing and retarding electromotive + _ _
[ . _ l MffdipdfiM :hhhsh% e j forces (emf's) do not overlap. If r = 1, o@h hhhhH the waves travel with the velocity of light; ' -v- - - v, - - - J. 6 ~ Fig. 3. Limiting . . l . N since the ritardation voltage is zero and , _ _ , the voltage waves propagate with a verti-
"'"" ""** "!NNNdNNNNY @hhhhhOhhh ldididit"M'
- which produce e l cal front, the forcing voltage is equal to
;) , ,,,g,,,,
I i', G0i, in D/a. For i,= 10, this expression will be recognized as the conventional A-System of simul. (c) 1a surge impedance. In addition to the (*"*5 * *ve5 .so y j .,u:'; ,
' B-Series - forcing **6%
- longitudinal electric felds at radius a, volte es, inser:ed 14 ,
from which the forcing and retarding prooressively in e A f v'60smt voltages were derived, radial electric sere r ired to ii 6 elds .which arise from the charges f* produce waves ^
, also exist. shown in (A)
I
, Before proceeding with the application C--Total vclieec to !
d of these relations, it is interesting te, be mserted L . . .! . 610 lt'agner-Stroke Current and the Vdocity of the Return Stroke ocronen 19G3
m.. I n,. ( Recteasurst (A) dj
.wr .n. ,.s n , ,r o rn . ~3 plane of reto re- b <
j l f_____-.-.]'__-,------- - - - - - - - - - - - - - - A--Current wave h i j i a, i f--- - 8 'y -E B - Longitudinal we -" i }= - y df l Fie!ds produced by > E 301 #' l current shown in ' ' c NT wave (A) AssociATEo wiTH '
- p i'.
C-for: ins voltage IMAGE CHARGE LON3iTv0W at. l , reared to produce QELgu Ao o
- N '5 es.rrent shown in ! $..,(
Ei sv (B)
, 30s, I Ty 2+0-v8)oz h m,
<%. e.
l - YH[ j:I'lf , at jt u. 4if (c) I-l, 7l I
.f-q i
j a gi . l . i ,e i u ; 1 i. i : i [RETARoiNG uEAsuREo VOLTAGE FRoM X, < kron wo votTAGE v
. % n(t. q )
= 30A-,nn(4 9v' '7. , )
kl(?
]L 3 3. l l
> l l '
a
.; , 1 !
i y!! : ' n's l ; i 10 d.,i 90004070 l3 when V, is inserted from equation 6 ist
- 30 r- In i,8 = rc(0.002)i, (17) *!I ,
equation 1-, e a
,.t j[2 3X g, 10' ,
1\in D- r,' or watts (15) P. - - 30 (r - c rs Ic the foregoing it was assumed that the effective radius at which the charge h j i which is. the power that must be absorbed i, = (0 002) - (18) might be considered concentrated and 9 : by the heaa of each wave so that the 30in g 1 - r8 b a the radius of the cylinder in which the ' E waves are retarded to a velocity r. current flows are identical. This is not i2 From assumption it is revealed that the energy required to raise a spark to a and true. The charge is concentrated at a blI 0 002) much larger radius than that of the conducting state within the time interval r=1 (19) evlinder along which the current flows. involved in the breakdown process of a 30i, in E 1 d' hen this is taken into consideration and b; ( rod-rod gap is 0.002 watt second per a l the radius of the charge concentration is Ji , centimeter per ampere. The power that represented by the symbol b, then, as ,) A certain similarity in the form of this must be absorbed bv the head of the traveling wave, to bring it to a conduc-equation and that of Lundholm Rusck (equation 1) will be observed. The shown in reference 9, a in equation 4 should be replaced by b. Following this [d(I [; t l j tivity which will support the current of change through equations 6 and 15,it can author does not wish to discuss at .is ]lj I s, amp, is proportional to the velocity. be seen that equation 1" then becomes ?. point the numerical values of the .am. ] The power P, can also be expressed, ; g .Qg, j t ere re, as that a typical value of D would be about 30i,8 ,!n I -rin- - tc(0.002)i,' GI) f'r, f ; p, = rct 0.002 )i, (16) 600 feet, and of a, about 0.1 foot. For -' "- Equating equations 15 and 16 these values equation 19 becomes and 9}5f; Ocroatn 10G3 It'agner-Stroke Current and the l'elocity of the Return Stroke till A1
b? 7 ,
* ~
current that rises from the ground; turn stroke has a very low drop (of the 1'
- 7. i, . _._tM 0 2 IU order of 60 volts per centimeter)- The re.
{ simultane usly, a corresponding current -( i 30 in - e' in c, begins to flow from the end of the down- turn stroke can then be conceived as a metallic conductor of essentially zero re. .[ . b ward leader toward ground, this current
- $ being supplied by a wave of current that sistivity that extends itself vertically.
5 The field along the surface of this conduc. Pr tresses Upward, draining as it does so Y Representation of the Return Stroke t r must then W zer . To pmduce the e cyge la d wn W Ge d wnward currents such as premised in Fig. 4(A), $ by a Conductor Rising frorn the leader m its progress toward the earth. forcing fields equal and opposite to E and i Earth But for the present discussion the re.um current will merely be assumed to rise retarding fields equal and opposite to E, y l Up to this point, the stroke has been must be assumed to be injected into the I ) fr m the carth. represented by waves of charge (and cur. circuit. This can best be visualized by i For preliminary consideration, both assuming that series voltages areinserted - 3 i rent) propagating along. parallel paths. Actually, however, the downward leader the charge and the current will be as. progressively in incremental quantities to j sumed to be concentrated at radius a. distributes a negative charge that is series in the conductor, so that the inte. ,C As shown in reference 9, the current wave j ' approx.imately uniform along its entire grated values of these fields are equal to
, of Fig. 4(A) that flows to the right from these series voltages. The values of J $
length. The return streamer or stroke O produces an electric field parallel to the
- can be regarded as draining this charge these voltages are indicated in Fig. 4(C).
path of propagation, given by the rela. It is dificult to generalize and to draw to earth as its head moves upward, or as tions in Fig. 4(B), trhere a is the distance ) 1 I neutralizing the negative charge by an upward-moving positive charge. There, from the axis of propagation. The field conclusions frorn these analytical expres-sions. A particular case is, therefore. E develops from 0 with the speed of light chosen for numerical evaluation; this is 4 fore, the return stroke is represented more and, therefore, outdistances the head of given in Fig. 5. For this case t is taken l :f realistically by a wave of charge (and cur, rent) that rises
- vertically from a plane of the current wave. The field E, travels as 0 3, a as 0.1 foot, and x, as I,000 feet.
]* with the head of the current wave and Of particular interest is curve D, which j 2
infinite extent. Again, in order to state consists of two components: E,, associ. J specific conditions and to obtain precise gives the voltage that must be inserted in
} ated with the charge, and Ei, associated j
results for these assumptions, a rectangu. series from the head of the current wave
' with the current. These f) elds are also x, for the wave to be slowed down to r =
lar wave of current will be premised that circumscribed by the sphere that expands extends itself above a plane of zero resis, 0.3 and the current wave to be rectangu. j from 0 with the velocity of light. The tivity with a constant velocity tc. The lar. This shows that for unit current a ,' r image current to the lef t of 0 produces total voltage of 900 volts (the value of i effec't of the plane can be represented by ' similar fields but, for large values of x., the replacing its presence 1,y an image charge this curve for x = 0) is required. Thus.
) instantaneous position of the head of the wave of opposite polarity and its associ. the power that must be absorbed in re.
] [' current wave, only that component of ated current wave as shown in Fig. 4(A). tarding the wave as determined from cir. the f, eld produced by the image current cuit conditions (that corresponding to In passing, it should also be noted that l which expands from 0 and is equal to E j i this representation is not accurate. Ac. equation l5)is is of importance. cording to the author's viewpoint, the p, . 900i,8 (:3, f return. stroke channel current begins as a The arc plasma that constitutes the re. i
*@ s (bem) -
l 9 i h - 180o- , FORCING WAVES f Ii isoo . Gk , ,,,,, l e
,a, soo-
, ..........................,_,.(.,--......... .
, * . me e,ao nee. ,
j , g .. t i i ,*o- ,o \ \, f?N- E **J t[,,,,,,,...,....,,,,,,,..i.', \ \. g .,i.n e..... ..
'ik. sw. .. ...m .. x. . ,
- 3. 1 S e inie,,.i .. si.h.n.,y f,sid e,..... i- .
n ..--.....
- u. v e., . -
5 y ; 200-200 soo 400 500 soo no aco soo t too ' ' ' N o
' ,'k e tioo 1200 13 o 0 .400 em
- t. . o 4 " I IN FEET
- y .
x - r . n .1 90004071 ti Fig. 5. -Forcing end retarding volteses in the return stroke after the head of the current wave has progressed to I f,i earthf a rectangular wave of current was assumed for the fullline computations and a current wave with a front of 1 dotted line computations (assumed velocity of propagation,307. that of light, redius, 0.1 foot)
; !j ? !I OcTonen 19,n f12 11'agner-Stroke Current and the l'elocity of the Return Stroke
s I o, n . c e rt os to to so e Fis, 6. Velocity- 4, , 9 d fr m e u n , o, [- ,/w/:
^ ~
'~1 - .* ~ ~J ~ ~ -'- l 30 for y- 300 feet, a = 0,1 foot, and
, / i 9 8, I
diNerent values of b i g3 w, . s. $h[ V ' s .c .-
<8 Y
- .p. l NE l .
" , * ~ ' ~ ~ , , ^ '2ow
- 3. f. 0 2 g7
^ '
L,, t: g ", , , . . 8o
///
,,,,,,, -ts d [fi . t oa .o -* I
,j 'a l :p'f .j l ,,,,,,,.# -"' -s [ I. p.
7
' ,.. f g o
o go ao 6o So loo I2o too 16 o i8o 2oo 22o too tit r 4YI STROKE cuRRENr IN Kit.oAMPERE s
- i, h' i ili Equating this to tac power determined and 904 watts within 1,000 feet of it. c = 3 X10", equation 27 then simplifies to]
by laboratory tests, as given in equation 16 for this particular value of r, Thus it can be seen, as would be surmised, that the energy absorbed is dependent i,
- on v -
(28) (%(h p a ' g. l~" 900i,5 = (0 3 X 3 X 108)(0.002)i, (24) nly upon the velocity, and is independ- In a j $ y est of the front. t or From Fig. 4 it is shown that the re- This is identical to equation 18 and dem- l,5 i,- 20.000 amp q age We recta $ar wan onstrates the similarity of the two circuits. k-is equal t For y/a equal to 3.000, this expression To determine the efect of the front of reduces to the current wave, the rectangular wave at
/3 \
30i, unit amplitude was replaced by a steppec' -r} In X ' i, = 2 3 x 10s (29) yy- - - wave with ten steps of 0.1 amphtude each, y 1 - r' I.. as shown in Fig. 5. The correspondmg , o v i _,s] ( 1 - 0,
+1 (2M e forcing and retarding waves are indicated
- If y/a is 30,000 instead of 3,000, s,, is only increased 26To.
{ l{f bv the dotted lines. It will be seen first where y is the distance measured back
.Now if the diference m, charge concen-that the amplitude of the forcing wave, from the head of the wave over which the q which corresponds to its surge impedance, field must be integrated. The energy tration at radius b and current concentra- t is unaffected by the front of the wave. that must be absorbed to reduce the veloc- ti n at radtus a is taken mto consideration ..
and the same simplifications that led to f To determine the retarding efect it is itv to r, then,is this expression multiplied necessary to evaluate the power rather bv i,. Equating this expression to P, equ tion 23 are included, the expression ; 8 than merely the voltage. For the case fr'om equation 16, for stroke current becomes the same as ; ir which the current is rectangular, the equation 22 with D replaced by 2y. Thus ; current to the right of x is zero, and to i L 30s,'\ h- rin X .
,s(o.
c cog) ( } ' the left it is constant. The power ab- ' 88 "
- Oy sorbed at the head of the wave is then the 30 d In2y~ .
}
[ y [/ y
~in7-t product of this constant current and the
- yg,+%qg- ,)\s +1 ~
f [ series voltage indicated bv curve D. Thus The family of curves of Fig. 6 have been
~
for unit current, if the wa've is rectangular, =te(0 M2k, (20 drawn from the above expression, with h the power that rnust be absorbed at this e a equal to 0.1 foot (3.05 cm), D equal to p instant within 100 feet from the head is is"i"- , X 300 feet (91.4 me ters), and diderent i o 690 watts; within 300 feet, 800 watts; values of 6. This particular value of D .. 40 m2) and within 100 feet, 000 watts. But if was chosen because it was estirnated to [ q( the front is stepped, the current to the 30in - C - == & a reasonaMe de dat wd od ; j left of x. between x equals 900 feet and .o VI-t' 1 \aV1-v/ +1. over the entire range of currents as the p 4{ x equals 1.000 feet varies, and it is neces-(20 stroke current attained its maximum ! sary to compute the voltage across each value. Further, it has already been y]i i, 10-foot interval, multiply by the corres. For the present purposes, interest will shown that the results are not critical ponding current in each element, and be centered on values of y greater than 300 with respect to the choice of D. Because l%p. integrate the products. This has been feet. In the discussion of e.rc characteris- of the factor r2 fer the term involving the 9 ' i done; it was found that the power that tics in the Appendix, it is shown that the quantity a, the e., ice of a is also not l [ must be absorbed between x equals 900 growth of the diameter of the are requires critical. The radius 6 is, however, criti- j j feet and x equals 1,000 feet is 554 watts considerable time and that, for the eur- cal. The distribution of. the charge Qg for a crest current of unity. Adding to rents involved, the are radius will be less around the core of the downward leader pij p this the power that must,be absorbed dur- than 0.1 foot. Therefore, unity under the is still controversial. Is the volume dis- l ibp ing the interval in which the current is radical can be neglected. Furthermore, tribution uniform within a cylinder at h F. constant, it is found that 794 watts must y/o will be so large that 1-r8 within the whose striaee the gradient is 30,000 volts : 'h ; he absorted within 300 feet of the head, radical can be replaced by unity. With per centimeter? Does almost all of the y; ,i 'j-. OCTolf R 1963 Wagner-Stroke Current and the Velocity of the Return Stroke G13 f 'l 90004072 . _ _ _ _ _ - _ _ _ _
*i
, Q' charge reside near the surface of this Fig. 7. (A) Geo voltste and (B) cunent for spedover W % , ,,., W ej tc
%' evlinder ' Or does the volume distribu.
t$on varv inverselv as the radius to pro- of . 9-foot rod-vod gen, D*ES3 MkYj N'k with (C) their [lL*.*d11 RG!i;iEG E
~
duce gradient within the corona sheath' '*8dh" MV@ r mj s-h reptots and insta ntan e ous GC EF#REN VE.L Perhaps none of these conditions is cor-p k7,ggr,y.ggrey p-j- i rect. However, for the present situa- gggg% # C..6
'A ! tion, some assumption must be made. b_ IlaLEiE in D U M The author has made the assumption - ""T "
'#hG%d
-4 that b for the value corresponding to -- M3% T:TFirm 3
D
]')
g 100,000 amp is 10 feet. This establishes point A on the final curve depicting the 7 mo j g(f g l l { relation between t and i,. It will be " 5G 'E" " 4,g l
*- assumed that b is proportienal to q,. sq 2400l *' ' 'dd -
{li From equation 2 it can be seen that b is '* If I l ! , i also proportional to ir/t. Therefore, to w l'
$y determine the point on the final cune for 5
! 2*
! " '5
-h b equal to 5 feet, one must only draw a fI !
M ['#'N! ggl I straight line through the origin whose *qy ,g l ' I g slope is twice that of a straight line n ;{ A through A. In this way point 3 was ob- S l8 , ; d I tained. By sitnilar constructions, the , 3 ' 9 i2co - -' '
$ dotted line sbown in Fig. 6, which repre- ,~
l po.ca I l T l sents the relation between t and i, and i i 2
'N l which incorporates the variation in b, ] . mo ;
w as determined. This curve is replotted ,E '
.Q ;
q in Fig. I for comparison with the curve
,j 5 E ,oo / '
I obtained bv observation and the Lund- t ; I
' 'j { holm Ruseb curve. The dot-dashed l$ { a ^
j
., ;l curve of Fig. 6 shows the correspending o- o d values of b.
Y l j l N -
-zoo I l I !
Relation Between Stroke Potential * ,2 s . s y i [,,,, , sgo,m, j and Velocity of the Return Stroke
, In Fig. 5 the series voltage !* required M.h . to establish the current i, is indicated. It knowledge of the factor a, which has been value of 6,000 volts per centimeter is i l
M i can be approximated by the equation taken as constant and equal to 0.1 foot. used and the final striking distance X. 4 But in actuality c varies with current as becomes j t; , V- 60i, !I n I (31) well as b. A more realistic assumption is d r b to let b be the value given in Fig. 6 and
- l. X, - 2 X 10' I - 2'2*'c:n (M) l let a vary linearly with i, being 0.1 foot N The radius for the charge concentration Nh must be used in this expression. Insert-ing i, from equation 30 into equation 31:
at 100,000 amp. When this is done it is f und that V can be approximated very X,=656 e, 1 - 2.2r8
" ( }
closely by the following expression: k,] d er 1 Table I also indicates the range of these h P = 1.2 X 10ar ,. . in voits , T 1-r 1n2 y In 2y P = 1.2 X 10' (33) values.
+ .
h, i 8 a 6 1 -2 Ocs h ;;Hc (32) This indicates that the stroke potentialis independent of the stroke current and is General Cotnments The main variable in th.is expression, a function of the velocity only. In
.4.
. In Fig. 5 the curve A +B+D represents
. except for t, is b. This factor can be esti-Table I the stroke potentials are in. dicated the voltages that must be inserted in series
< 4 ,g mated from the curve to Fig. 6 and i,
,t . for different values of r; it can be seen in incremental amounts at that instant so t.b obtained as a function of current (or that they are of the order of tnagnitude that the resulting propagated current t ,i -
velocity). This procedure assumes a wave will be of unit value and rectangular estimated by other investigators. f 3d,.) According to the stroke mechanism in shape. Previously these voltages were i i ']! f' theorv proposed by Wagner and Hile- arbitrarily divided into two parts: (1)
- I g man, the length of the final striking dis- that from x = o to the crest (1,7GO volts) tance as the downward leader approaches was regarded as the forcing voltage; and
~'y I In[I,r.s la loFyoits in IIers InI.. the earth is equal to the stroke potential (2) that between the crest and 3,300 feet t.thh. divided by the crit _ieal breakdown gradi- (about ~ 1,600 volts) was regarded as the
.b 20.000 :s.o. 30.. .te4 ent, which is about 5,000 or 6,000 volts retarding voltage. So long as these net
*1 NN .$ .y per centimeter. The smaller the final voltages are maintained the same for any
,. Iw.ooo .. 6t4N
- o. .' ..211. .n4 striking distance, the smaller the shielding two cases (but not necessarily constant k $N s.[ g j angle must be to provide adequte shield- with time in either case) the two current
' .u, ,' [h. L.. ing. Therefore, to be consen ative, the waves propagate with the same velocity. ,c; ;h 614 Wagner-Stroke Current and the Velocity of the Re: urn S:rcke OcioBen 19G3 i l 90004173
l i The canne.r in which they are distributed Regarding the downward leader as a mation is known concerning the charge ! I conductor of equal potential along its distribution within the corona sheath, not L will afect the wave shape of the current. In determining the retarding voltage in entire length, the charge distributed in the only during the 6rst few microseconds equation 25, the retarding field was inte- charged column will be larger near the ( sec) of formation, but for longer times grated from x =y back to x-o. It would earth than at points at higher altitudes. during which other factors than those ', 1 have been more accurate to have inte. This should result in a gradually decreas- dominant during f ormation may be active. grated from x=y back to the point at ing current in the return stroke after its I, which the curve .4 +S+D crests (or 700 initial crest has been attained. Since the Conclusions [ feeth However, it was shown that be- velocity of the head varies with the cur-cause of the logaritlunic nature of the rent, as indicated in Fig.1, the velocity of While the lightning stroke has doubtless J function this limit is not critical. the head as it rises should also decrease. been idealized and simplified to a con-j l While from a mathematical viewpoint Also, as mentioned previously, as the siderable degree in this analvsis, a number g, , it is immaterial whether the charge trans- length of the return stroke increases, the of pertinent conclusions c'an be drawn s ported from the leader is regarded as a voltage drop required to supply the ther- within the purview of theselimitations: negative charge lowered to ground or as mal energy conducted and radiated to the [i a positive charge rising to neutralize the surrounding air no longer remains negligi. 1. The general form of the modified negative charge, from a physical view- ble; it also contributes to slowing the equation relating the stroke current and - point further consideration raust be given to the problem. If, as the author be-velocity of the head of the return stroke. Furthermore, as the length of the return the velocity of the return stroke, first suggested by Lundholm,18 has been (g , lieves, the charge on the downward leader stroke lengthens,its surge impedance also verified, using as a basis the energy re- ;j l ' exists in the form of negative ions, the increases, which tends to diminish the quired to establish a spark. The modi- ! mobility of such ions is low enough to current. On the other hand, while the fied curve is shown in Fig.1. }; , prohibit the expla=ation of the formation voltage drop along the downward leader 2. It is shown that the velocity of the L; , of the current as a Bow of negative ions. is small and negligible during the initial return stroke is dependent only upon the Ut ! The current must rather consist of a dow stages of the return stroke, its integrated effect over the length of the downward energy necessary to make the are con. ![ l of positive charge drawn from the earth ducting and not upon the rate of rise of that neutralires the negative ions. If the leader results in a higher potential of the downward leader near the cloud than the head of the current wave in the chan- dl ' return stroke is viewed as a highly con- nel as it moves upward. l near the earth. Thus, as the return h ducting extensible conductor that rises from the earth, this current must rise by stroke approaches the cloud, the higher 3. The potential of the downward ,h .j induction. As shown in reference 16, the forcing potential tends to counteract leader as it nears the earth and the nnal gl magnitude of the current that would flow somewhat the erTects just enumerated. striking distance are functions of the d in this case is determined by the relative Several of the parameters used in this velocity of the return stroke and are .jh radii of the return. stroke ch'annel and the study are not known precisely. They given by the following relations: qj charge location. The more nearly equal vary with time, but specific values still e {i, these are, the more perfect the neutrali- must be assigned to them. The author 1*= 1.0 X 108 g (33) p, does not, therefore, claim a precision j; zation and the larger the current. It is probable that in its upward traml the that does not exist. Fortunately, these # (35) U;i X,=656 1 4 h ' feet return.strcke channel branches out to parameters occur as longarithms and so /: collect by induction a large portion of the they need not be known with great preci- i ionic chari;e. The core of the return sion. The precision is suihiently great to Appendix. Characteristics of f4 stroke does not increase from absolute rero, but from a value existing in the core denne the nature of the phenomenon. As more data become available, appropriate
$ parks and*Arcs M Si of the downward leader at the instant iti corrections can be made. More informa- Reference was made earlier to tests
- in N which the head of the return stroke tion is desirable concerning the simultane- which, from the breakdown of spark gaps, p reaches that particular section of the aus variation of the velocity of the return the energy required to establish a gwen are current was determined. These data jlp downward leader. stroke and its current. And little infor- were for rod-rod gaps 6 feet m length.
39 . I While at the time that these tests were v! made tests on gaps of other lengths had lAl 5 , been made, the ene:gy had not been com. :j puted for the gaps of other lengths. Fig. 7 [ib
. shows the form in which these data were ji[
collected and represents: ( A) an oscillo- y. m o O gram of the voltage across a 9-fout rod-iod I 0 y gap of such a value as to produce complete i . i. y , c' breakdown in 3 asec; (B) an oscillogram ,;' u F o= * - of the current through the gap; and (C) fi; y . A a replot of these two records, together 'ji
; with a computation of the instantaneous 'i product or power. The total energy ab- , N,'
i2 o . sorbed in the gap was found by integrating i(b g this curve, whicn was found equal to 3 06 ] g ' watt-seconds per centimeter length of gap. , w 0 ' o}#
~ ~
Fig. 8. Esperimental evalue. The integration in this case was continued , l *-9FT - only to the 6tst zero of voltage. The value I tion of energy led into rod- of the current assigned to this test was the ;; {i . o4 . tod gaps of b to W oot value obtained by extrapolating the smooth g, o aoo Boo 12co 16 0 o 2 coo 240o spacings during spa #kover as portion back to the instant at which the ', , Cl#RgNr IN aMPgRg5 g function of the Anal Current Voltage became zero. This value was m Wagne?-S!!0kt Cu??tn! and the l'elocity of the RtWn $ltokt 010 OCTOBER 1903 l!
a 1 o properties of the arc. Mayr" show ed
,M D that fut a steadv state condition, the
]r {gg current distribution' within the cure of the are falls off from the axis as a quadratie Bw2 j:v . E'[
>3m exponential function of the radius Energy losu is by conduction, although radiation 5 350- 70- 7 e also plays a part. The arc core is sur-d s .e /cm/ome $E rounded by a sheath of air at high temperc.
" ,6 - ture, since at the core boundary the tem-h d
2 XO Y , E C""*"'
-" * * * * .. .. . ----- 0 0024 perature is continuous. To determine the g5-
-0002 energy or " thermal content" that must
{250- g 50-
- be injected into the arc when the current 5 o g ,
j > 200- w 40 8 4- / Diomet er - 0 0 o16 rises slowly, the energy imparted to the
- E 15 0 - E 30-( ,'*
-0 00r2 surrounding air nust be included. Yoon and Spindle," in observations with an f3 g 3' ,
8 E are of I amp, obtained a thermal content g4 E 2- -c ocos I} o 100- 2 of 0 0025 wattacevnd for air at atmospheric y o. 5,, E 30
,/ -o 0o04
#***"#'
- T
'. o. p.
fo **
- incremental increase (or decrease) the thermal content should vary exponentially 0 to 20 33 40.,,, to so'o with a time constant of 85 usec. This TIME IN psg0 slow phenomenon does not include the very rapid effect associated with impuls.e cut-rents. Fig. 9 shows the growth of the ,
h diameter of the arc, as obtained by a photo-M Fis, 9, Charecteristics of a 68,000 amp damped oscillating evnent arch that rose to crest in graphie process by Nonnder and Rarsten. 1 19 piec, length of sap,51 em More,recently Allen and Craggs.u photo-t
. graphing unpulse currents that rose to a crest of ISS kiloamperes in 7 7 usec, dernon-l .
strated that the initial velocity o! expansion 9 y 1,590 amp; therefore, the energy becomes mately within the first 5 usec. It can be of the are w as 1.4 - 3.1 X 10" cm per V equal for this case to 0.00192 watt-second concluded that the energy input per crest second, as compared to the velocity of
~
per ec::tirneter per ampere. ampere increases with current. It is also sound of 0 34 x 10" em per second. It Fig. S gives the energy data for gaps 3, clear that for long times the energy per seems reasonable to conclude that for the 6, and 9 feet in length. The data for the ampere is well in excess of the value of rapid changes of a few microseconds in
} 6. font gap are those given in reference 8. 0.002, which again indicates that the duration, the surrounding air does not in drawing the mean, the data for the velocity should decrease as the head of the participate in the heating effect; howes er, 9 foot gap were favored. The slope of the return stroke approaches the cloud. it does participate for slower phenomena ti straight une is equal to 0.002 wattoecond Wagner, Lane, and Lear" also measured and phenotnena of longer duration, such q
per nr,timeter per arnpere. These data the drop across high. current impulse ares. as when the arc path of the return stroke
; covered a 'eurrent range up to 2,250 amp The gaps they used were small (less than has lengthened as its head approaches
; and channel formative times from 0.4 to 0 inches) but the currents rose to crest in the cloud, or during the dissipatioo time of 65 usee. much shorter times. Fig.11 shows some the stroke between cotuponents.
l of their data reduced to energy input. Frorn this brief and necessarily incum. j i While these tests indicate that the
! potential across the are path drops almost in judging these data it is necessary to plete discussion of the arc, the additional j '
immediately to zero, a drop of, for example, bear in mind some of the very complicated luformation obtained from other than gap l 60 volts per centirneter would be dif5eult t
' to detect. A drop of this magnitude is equiv alent for a duration of I usee to
.1
;l 00N5 watt 4econd per centirneter per 000*- "D 1
l ampere, or 2.5% of 0 002. For the pur-
- l ! poses of this application, one is interested i
f in onir the first few microseconds after f ooo3 4 crest current is attained; thuefore, this ,_ j conti:uous contribution to the energy can I" -soooono 1 be neglected. It can, however, be a con I 1 (' tr:buting factor along with others in ex- 0 002- u s,ooon,
- 1
~3 plaining the decrease in velocity as the head {d8 33ooons
; of the return stroke rises toward the cloud. %,oo, ;
k[g,i, Consideration will now be given to other avaibble data from which the energy input
; oooi .
g / j l [ [i to bring an are to a high conducting state / [ k,l can be determined. Norinder and Karsten," by means of a combination of 0 Tj p impuhe generators, produced currents up '00'
.,j g to 102,030 amp through gaps of 25 and 51 iorooon 1j i- em. Fig. 9 gives their results for a current go.
3
' that rose to a crest of 68,000 amp in 19 , -seooons q i; us ec. The current, the arc drop, and the y so- gog, 4 instar,taneous power are in the form given W 88 C#'
by the observers. In this paper, the do-i 3
- rsooc ris authar integrated the power curve and 3 a
s to-divided by the crest current to provide the i[ni
+
cune plotted by the dotted line. Fig 10
- o x so so shows simDat information computed by the 3 io go 3o d bl authat from the data of Norinder and = - 2o -
5
'!. Karsten for different impulse currents.
fig.
- 10. Energy _ao -
44 It is di5icult to draw direct conclusions from these results, since the primary laput into oscillatory [(i t intertst of this investigation lies approxi. evnent arcs '50' TIME IN MSC tH 11 616 Wngner-S:roke Current and the Velocity of the Return Stroke Ocroacn 19G3
m _ _ _ _ _ _ . _ _ __ _ _ _ _ _. L breakdowns, while not contributing to a Fig. 11. Energy g ! more accurate evaluation of the energy to input into develop- $ 00 2- 2,2co,, 2esoon a ! use in the velocky armula, does not ment of an oscilla- g contradiet the vahte determined from the tory current arc (first 3 ssanon breakdown of long gaps' number represents me gnitude of first I' o.cos. rso^ <2roon a f current crest; second P., NN number, time to Arst 7 '8'00/* *
- crest) q p s op 20 00/24 J
- t. Inpucso Ova a vottac ts ow T aa ms uission f
Lamas ano Tusta Baan.smo ow Tus Laourwino -g PsaromXAMCs oF M so!UM VoLiaca N sTwonts. g J R. Lundbotm. Duplic Goteborg. Gothenburg, y j mm . t' Sweden,1955. E he iW 1 I
- 2. Inoceso Ov s a votrao s. Sue c as ow T aAm s-Z k '
O , ,
. . r, nassioN LINS 3 aMo Tasta Bsas.INo oN rRs , ,
O 2 3 4 0 0 Lacurnino PsaFoEMANCs af Mspiou Voctacs 8 9 Netwoass. R. Lundholm. Transactions of Chab TlWE IN pSEC k mars Unsseruts of Tec4sology, Gothenburs. Seeden, vol.183.1957, pp.1-17. (,
- 3. Inoucso Lacarweno Ovan.Vooracas on t Powan.Taansassuou Lawss wtru SesetAL Rs'* core or channel, and a corona sheath that gradient rnust be of the order of several I
$s, ,$ precedeA and surrounds the channel."1 thousand volts per centimeter. If onc j how".No[tacs NarEo s Sune ue assumes this gradient to be 5,000 volts '
scaoss of she moyez 1mssu=4, of Tec4=ozoty. Stock, Photographic evidence, mostly by Schon. J boim Sweden, vol. t20,1958- land and his colleagues, does not indicate per centimeter and assumes that the leader ,
- 4. Iwoceso Votracas on TaANSMiasION LIN as, lhks. The corona sheath has not been length is 5,000 feet (150.000 cm), the drop C. F. Wa gner, G. D. McCann. AIEE Transaasons, pbotographed, but has been irnplied to be along the leader channel calculates to be h vol 61.1942 pp 916-79. a so-called pilot leader. The photographs 7.5 X 108 volts, quite beyond any modern 1 5 Paocassarvs Lsostnruo !!. B. F. J. Schoniand, show that the conducting cores, which are estirnates of the cloud potential. Further. 1 45y foUo ,LInd n t permanent, are formed and extingutshed more, one would expect that the boundary )'
[o's ad vL in a very short time and are reborn about of the resulting space charge (or corona) 1935, pp 595-625 Wery M usee, each time progressing further would have the form of an inverted cone )
- 6. A Matson or Est:KArtNo LicMTNiwo Psa.
roamance or TaaNsMisStoN LINas. A!EE Com. toward the earth, possibly collecting posi- of enormous radius near the cioud. Thts ; mittee R eport. AIEE Transadious, vol 69. tive ions produced during the progress also is contrary- to all modern thinking. 4 The conducting cores to which Mr. Hagen- ? l pt, II,1950, pp.1187-96. of the invisible pilot leader and other
- 7. FusasnaoNsraWrs, ZeNorQNasN ANo positiVC dtbris in the Corona sheath. I guth refers are formed very quickly behind / l WawosawsLLs. M. Toepler. Archiv / mar Elektro- am not aware of a photograph that shows the steps, so quickly that their velocities ,
f
"" a continuous core of the downward leader. approach the velocity of light. Such rapid v$1 192 . p $4 8 Macnautsu or Banamooww or LaRomarotV from the Empire State Building 8 with only take place tf the channel already Cars, C, F. Wagner, A. R. Hileman. AIEE T, coasou pt. tit (Power A pparatus and several portions where the core appears consists of a good. conducting path. The p i syu,mo vot so. Oct test,pp 604-22. ter be continuous. Possibly in this case fading away of luminosity mercfy rrflects i
,,,,,- 9. Svaca !=rspance ano its AppLicartoM ro the conditions postulated by Dr. Wagner the reduction in curren t. Behind that . , rns Licutntwo Srnou s. C. F. Wa gner. A. R. existed. This one perhaps was the notable part of the path which produces a trace on , Hilernan. Jhd . 1961 (Feb. 1962 section), pp. exception that proves the rule. the photographic film there must exist a 4 j In connection with the relation between srnaller current not suf5iciently photo-16I . t 1961, p 67 8 "' velocity of the return stroke and the stroke graphically sensitive that feeds the forward R. Hie an progress of the leader. In this region the current, it would be interesting to know n where the field data of Fig.1 originated. eurtent should be almost constant and of ;
$rANC&a sa TH1 . E r! Cl t.!C R ?
Has it been possible to correlate Berger's such low value that it does not record on wiwo Cunasur Parns. H. Norinder. O. Karsten. Archa for blatemetsk, A stronomi O<h Fyuk. data on this basis? the photograpbic blm Sweden. vot 36A. no 16. pp 1-48' Mr. Hagenguth also inquires about the 7
- 12. Aac Daoe Deatwo TaANstrioN PaoM SpaaK REFERENCES source of the 6 eld datq in Fig 1. I believe h.Y.$a'r. ANE ra so [oYs e !! ,'l 1. Rs LA rios Sarwesw Srsors Comment ano that, the text explains this in detail. The f, A p pa,atus au Synemo. voi. ?? June 1954.
VsLoctu or tus Rnvan SnoKA C. F. Wagan. ortginal data consisted of the 19 photo- 7 1 pp, 32 47. Thornton Butterworth. Ltd.. London. Enstad. graphic records of strokes obtained by . Schonland. Malan and Collens. Wagner v 13 Bertaics sus Tsaonas oss sTatiscusM UNo See reference to of the paper. and McCann then analyzed these data, oss ovumiscus= Lacutaooax. O. Marr. A rch, 2. h fear E;setrmeskuk. voi. 37, no.17.1943 pp 5%- 3 Lscnf utwo to rus Eueras Stars Burtorwo- from which the velocity curve in Fig. 2 608. Paar !!!. J. H. Hagessuth. J. G. Anderson. was obtained. The stroke current curse f{ ALEE Tra nsacane r, pt. !!! (Pots're A pparata s EE m n' nm' 14 A Sruor or rus DYNawrc RsaPouse o,
^"8 1942 PP 441'49-M ' o b ~ '
Apes su Vastocs Cassa. K. H. Yoon, H. E. Spiad's. 8 84 878'*
- 81 L Il assuming a singMalued relation between AIEE Trauserwar. pt. Ilt (Pows
- A pparatus end Systemo. voi 77,1958 (Feb 1959 sectioni, pp the stroke current and the velocity, the 1634-42 6 eld data curve of Fig. I was obtained. E:
15 Hson Cuansvt Spaan C n a n u s ta. J. E. C. F. Wagner: Mr. Hagenguth has raised To be more specine, any single value in t' Allen. J. D. Cra ggs. Bruish Journas of Appiud a question concerning the evidence upon Fig. 2 on the percentage ordinate will n F* *'u' L*"d'" E"8'*"d 5 1954 P P 446'53 which ! postulate that "the downward provide a single current value and a single !! leader consists of two parts, a very thin velocity value, plotted as constituting one good. conducting core or channel, and a point in Fig.1. f'>i corona sheath that precedes and surrounds Subsequent to the preparation of this j the channel." There appears to be little paper it came to the suthor's attention that 4
- doubt concerning the presence of a corona Dr. Moller.Hillebrand had plotted similar. ;
Da15 CUS$lo f1 space charge. The only question must curves between the velocity of the return ;
?
refer to the presence or absence of a good- streamer and the current.1 J. H. Hagenguth (Ceneral Electric Com- conducting core. The head of the down. l pany. Pitts6 eld. Mass.h Much of Dr. ward leader ruust be fed either through a REFERENCE Wagner's theory of the lightning stroke is channel consisting of a thin are plasma Q/wwwh h h v' Tns Pnrsses on tus Lecurutwo Di4CHanGB. h f Vk [( based on a leader mechanism:5 8 he has whose voltage drop is about 60 volts per L. d postulated, "the downward leader consists centimeter. or through a glow discharge. (*[ ,",'* *c,',,,,fy',]; g'2 1. p b of two parts, a very thin good conductirg [f the latter is true, the vertical voltage 232-s9 '! 4 OcioneR 1903 Wmer-Stroke Current arid the t*docity of the Return Sroke t\li i
, e e s
...s. v . ~ . ,
mo ygp., y D D O D'3-no e 1.n 6 n ,
- g. ,. . , . .. - ,.g. u g.-..,.-
- C ", b b. .$a . I '. 4 L: L . s i .., s1 t+6- .'
E.
M, " 4w . g.a=1n . .
uW (e g .jgu g.u. ;= a L .1 % wb b s tc.
. . .L.r ', ., . m. . . $, ,u ....,h. C./ , t,.. ., r-...- ..4
. , . , .... , 'h'"
' . , - . b
- e- p' ' .' ;II. ' ' ., L, ' ,". . . g * ;* I , A, .'M-,
' * * ~,' ',? .' I - ,, *, I '. * * "e* I ' -g k,,','w' "
t' - ' - ' * * * * * * * * '
; * .> :V , .* '.J d .";'.ri'* CS of the ::reke Tb: m m:;r iaiorm. vi :F' MJi;. ' ' n -
!.n U. n . r. . 1. w. .- ..-o. . ., . .v. . i ,. r. ]i . , e ...., e . . '. ~ 3s ,, .. ;~ t. or d. a. g L.. ...L. . .g J a .
r .
. . t. .m, ._..,
.,..., :... C.
..p.
..b . , . .. _ . , .L . :, .. g .Q 7.. ',,. .;..7
. . . . . t 4. ,3 . n .,. 2 3., . ~ .g. .- -
. . , i.
. .. .-., . . . . . . , .....,a. . . . s....n g . l .. , t. . c-
- s. . O h'n....e : , ..
.u ..
. . , ,. a ,
~'.l?'. '
.b ' $. ! i a r ['i ' " ' d., m;;2:'rdO]!03l 3sJ iiiNkr (O it.
. .. .. .t.
a ..n.......-
. ,, . ...- .. ,.. ., .i.
. . p-' .W. . -
T .* i eo . r "i :m. .
- 7. . . . ~. . r r. - .
~'-
.. ..j : , to, s a.n .i .d' ..;.
.. .. , L' ; Sa.... t. . . .
, i D . i ,
.. ..a. 6 u 3. .- . . .,a..
. .. ".-4.. .
. . .. . ...1 i
- x. v. t e .a win .:.; ; ; w::v. th:.* the 3erge's d .tr it.J.ea A; . u 1 b :ce b 't : emc;. % d anJ on ':s cours: hi-S s alue; one stic'.0 v.d ce n ..' ' ia cu< <
. mci '.h ec.rt ~F vr.:t ran> ny of strokr. cre ! WOO crar eres. :.e. ' o s e A.rveu u.
. r t i" r:: t* ir, thi.i
- .e.- l ewer a re.c.ative of om en cwfr;ner.; keeni t.m.r. 8 c:1. -
_ ;. ;o ::ar. " ?. re'c: Ov:nndly deve!ons t that originare fram the car 1.an "." % ;s. e
" he u% h.
. !* mf./ CC:hh Of N.I' 000 Cr C[ 3' VCr all hUi!00 T : !? O. f .M L . ' J "" C .
C 0 0. .' :."' - 4 32 "ry? V0'OCl!yl Cf ti".e UC.I.W; hM C c!:"" ? r.;t.' C 'i': . i , i d T.c ' L,.y . a-g .s g.. g H p ., p tr.y. ) ...w.
-.j...,c
. .. . . . e.. { . c . b. . Civ..
a . . . .
- , O.F n. . e g 's .,,. .i. 4. p o. . . , ,
o ..). . . , , ..i
.m c. ". gl r .
g [ , -.' ...... i. t-.. g. .,,. .w. ;7.
. . . %n .,s%, 0,c u. t. ,. ... ... -
\ 5 '*""' 0h'K Ch .' c ^.r! *' Or 3 f rO JC*I C r'.
; h ir6SRt0d ' Dk?
-e
. 0. S..i r . '. ' 4...
t.. t . .
.: .<:.. .. .,. . .;.:: .t... .1y,'.,..,r 6 .
6
. Ff ' ~ ~ * ' ' I' 0:rul O r f ~;'!7. th; pr0)C.!CN *.9 I'-d O @'OIWC W C w .
i J ';" .
'i..:2L**
4 ,
}5 f.ehv W' T ' t. 3.'.\' a e b. ..,. . . a h. . a sn f* I .. s . ..-. i p . r. . j '- l ",- -"3. .
sl . 6 . C L' V2 t . ) .* /a . B ,. .. l '. # d
. f..* , "I . LP.
.*)
. ,.i\'#'* ,e.*..i,.T4* .
^
. ii . r. ! . ' . VC ' J' .; 7(' . O I%2 Q'h ;UT. .g 3, , e, .i ;r to '
. ! a ' ' .' .'.'e r JO r. ' ! .# 6 ". O *
.t- .';" " ' '
I . .*,["",
,.,.i p- ,,',.r #
c
*) % [' ' (Ie 8,[ p g ,.
- f 8, g[ . j . #g, j gg
.' (' I T " ,, " ' ' 4"j$ "'8 ),
- g"'*
C .! ' . , (,"
- Q 3 ( 'l d
*} '^ .{* j]'. g f* t} P
{' ' ' l' . ; .- T o r. ' ;Z :.s 4.1 , Mim ; m i .er Ty n.-~ h
.u4. .
.....y. [ . . ., .. -. ,
. 6 . . ,
6 . . . * * . *
-4 .E . -
90004077 i
l OO O - l D D 3-
. , D f
Oc c b. . ._ t
- g. u hcre : is the curr':nt in arns: :s. v . .
~ !r.
l o
. ,<... exFr:ssed as a i,rn:u.on c:. ' .,. t c , e. m .
- n ,. oa . i I'~ : the charr,e in caulcmbs par em. This xpr:~m tr.n
- : .=---
w.,
.-ely'ro cives a valee et, 6 y 10"' coun..rn s m .
~"
r" ~
. . I
!. ,.I - 4 similar rearc.ning. the Icader curren:, d me . ": + : ,,
- d vor 1
*C ' l velocity of the downwarc leader b; 0.001 c. 't
. 9 !
- ~1", inv0rscly prcportional to 'he N J ves.':.=s. .$r - 0 +-
1 arnpere; Although the cen:ral cenz.:: Un.; ccc: 's
~
.>.'. L. p.. l
,, . %. .e : not actuauy be, a cher"ed on!y aa are :m. la - ;
nro"ide the nee.:ssary counctmt." h:2.: r i W
.: t
..- c ,; s . .
l
'2 5a QM CMcNA M & have sh.ma ' hat v:n .m are havie . :. Ju..
1
w N./ SHEATH .
between 50 and 1.000 a- peres it d r . . c r - . . . . . ..
! ::.: h. s hed, the drer ae:rea< s : rom an :mm ' mp j cMNNEL vaiU0 to 100 vi3 pr 'tn at ' n:IJ.im e'.H %' &
l '~ m [.) Lit:r 3:aHe Wu: of 50 vate pe :m ta ..i. - . j
'c x r ._ );
__, seconds. Thq 20 found @ ,t + abm.'.- l i FNU s.wp UO-amper arc b :10.:: 2 7.ir.. I I p' .,, For reasons rim in reuren:e i, tb n rmt .E B
; w l a.stociate, A. R. Hi!:mm ;.av e S.;e n p r:',-~ ." .:
i W .:- Q r p --._ '. a
- i. aswme that the c,aart: ex.r t".w.o : :n t,ac s. r r' -
)
o:>, L.g .- Q lead.: var:es invc scy as 'h n dia dLtc = : " cP l t ,j---=9 ! t':c: c:n'nne! exe. Fig, i. c.P.:n #- m i; . .:v.
-- a : " ,l- ,-- aL,,- ---: q i sao.. -s a.m, , j...,,,.., .m. aa n . e.. i. ,.m r . ),, ..., s .
e.
,r' ! .
l ,I.--<.~~.~.'"g,,t
. , i entC!an: "1th 'ha hri Oi F: J i' . J . " '.
*:: .--s.,'i. f g.} io 'H ' .' u ' .'. ..* b. '. ' '. '. 4. ' '. ' ' '. '"' . . '. . ~ '~ ' '
g . t . . l J- 1. . r* r r* '*--* e a Lil (l,' <*i<'* - . i '
,i. >- - -- - -
_ - .; ,-g.:. _- \ r. I - . .' . .
. .a i ?,,,,~,,, C ,,,
.to ,b (d'Lt:ua"" ;,.n[ '1de *.W!.'!.
'-" ~- ~~.% 4 b.A. s ,, c.w, , ,. . . . ~. p *:n'....
__ C o ~- ...,.,.- . -
.. , , . - ...~.-'... _
. . ,.. .. u.
i _. - -Vl . g { ' tauon, ih<: .m i e r "u. ir : .'. -
- m ;r .
M. 1) ' . , - ,... q . _. . - : d's...,.a q n; . . , s 4.. . .a. , .s-} ,, . e., .
,- ,,, *) v. g n
= j , -- .c,.,,-,, l ; .
ab...-e... . Jt+3
!f bra C3 D iM OU . L I 2' .' " N, '-
i . Lg ,
' j.; ,J/<"s- :: , '
y/ e
,"..L.i '...,
7/ (.a ;.1, . h -. .. .. ., ..c ... :le bA r., a
;* i o., ,--4 .
4 ..,,
. . , a . .
j *
*.m a# ,Ls ! ,3. 2 r. ' }. ., j . =. 3 ps. .- ,., m. n 3 *p. [ f.g. <
1."..
~
.'." l l
6 t' o .
; . . . ..r. g s. .4
-.s. ..,o. ..
gj ,..m.. .. , s
- n. - ..
-- - V i ......!... w..a3
... s.l .u
...a. ,. . , s
-~ a
-- ~ 7 , . , '%.,...
****e- ' * *' Ep's.
,,.',,,,,,0',,..., Is . 4 ,
e r. s g y, r,' 6,',.[. , U U,", 7/ .
.i..
- i. . \. ...a.
. d.l r.".V'...
..i .6 u ,- .- 4 ; s. . . .6 ,
l,, . I.." W ! . r e-o ; ' 6:e 't. un erc;n :a sHaQ Jt t t.t m.v! . ,
. mo.-, I n g .t.. . u. ,'t ,.h. !l a d. , .
.t. . -.
r.d ;, e c ,ursitd u t IM he:: :, Of tri:' Cntm* . JW" .I~n O f t h* , 'l $ ,j 's,.'* p m . , 7 '. i ,' .' .' .. . . ', . . ., . a p ' ' ,' -
=
it.i I n f 9 n') ' 'Hi a "rn t' n t' CPfl! ' ' ' * * ' ' N .h UU nd 0, '!! . my .h a~ , t ,, ,st . .a i ._...
, u,,, i .r: ,. c. . . 1 ,.,.N. . hat .. .. ,. C. .. .
- m.,t r O,.
m tFat in; corabion- vpr d d- xr- . i,.... , ratp. . i.,. v.,i.- .. ..g.,.. , , . ,. . _ . , , n., .... .o. . . .h., o ,, r o r , . , m . ., n.j h. .,
. . i..
- s
". ,b .a. .,p. n.. . al. . n . .#?..'..,,'.'.' " "#?.. "' 2.
enQui ..li c:rtp abrt .~3 w.hs per em. Tiu char;;c
',.. p. ,
I'* p.,.,l,.- .... g... u -. u .a
.3,,,.,.,..I.
,,,.I
,1'
, .- li,.. .
. . . .,. . 55 Oi
,3,.
.. . s. . .> . v . . . , ,
Ot .., r.,.
..,,,.;'.J m .,A. Q , . ..a.. ., h, n., . .... t- . ~.
'l .,.,,4 6.,.. H .,, 'I t ,4, 3., b. .. . . O,.3 s. .m t tr :.
..4 .. , . , . . . . a*
a..c
. i
,->,. Lc; a. ... . ,. , . - ., ' ) ri t. ,2
.a m.,.,,4..,.,. ,',,...+ ..-
.%s n
.. n. .t...,..
6
. o
, n .. 3
+, m. a . ... v!,.
4 ,
. l
'J "., '. .g... !.'
" a l ".. " . ,, .,
" " ' .A.~' ' ~ - ' '
rew rn r+ . o h:rmed by pregickin opping arJ 9 .. :r i t, dinrDmed charp to ea th as its ,. . . . . '
..3.,.,.,..y e... -.g..j...,i, 4 7.,. l ...,wg..tq,. . .A .g.3 .3 p ,.%
. . c.:a i . Da..:rr + . e. er ,~: w . e ,.n: , e: ,
l .. .
- .g., e p 3 ,.g. , ,
..< . .. .v .s. 7,g 3. , g ;, gg,; ;. , ,e 7 37,,. ,
SI".n! ld'l .." 50"U? C ar?. is inVOE,'Od in !!'? [O?tila- - - -- . ~ . .
.w. i _r .i . . p'. .. j .;n .v. j:au. ,, 4. ,4 b.;,a.l
. ,3 : , . . L.. e. c. :s C. .- = 0:r f m f a t CI
. . . . . i. . . . , . . .
- r. . :. l ,., s. 4 ..p... .,r! w...,,.
, . . -. 6. ,r ,
hv ! J '
'T
. . .u.. it . Jir*, .J.. ..w
, , , D. t L, .,! 3 1
... I
- g. ,n.. ... .*,. ...* : .l o , t. . 2. , e .uL.i u ,s1.... ; g. 6", t,.. , . , i) !!L.
eN'. .0^ S
^ *$ E3 '. I'. I ,C$c u [f C'.!.[) retuIlla 3 3.c : r. .. e :.s.t vr.:r of 50.G00 ampres. e 8
', ll .: 1' -
; tr" - M, % ":'ac.t' of th tip of Q:
'su rm.u -
.,.sn,- .
.- (, .r,.. b d ' " - ' t! '
.n.6.r-7 '
.s...a ,. .J em .,
s.6,- 3 4-
. . . . . P. . . .:. t . , ,
. .t . . , b
.. 4 .
yy. ' . . s
.ri 'W g - ,, gg e<*.*
---, ,, . . , e, .
. m. , a .'e c. d
. . l l
.:nawiza..g: of Ge surge imnedance of the return L t. e e c / reamer. wE IN -5 o a R According to the records of the electric tield uSIC- .
l remote from th; stroke. mad: by Kitagawa and l ' !, wti !, i A:obayashi' and also by Kitagawa and 3 rook
- the .! '. o=
i da: interval between st:p: as the leader approaches .' i, e, !
, ;. , i
- nc . earth ben.nes thort:r and shorter. While re- !
- m i..eoB : , .
, i., ', 1
~
- ca f ;r: /.g; carth. the rtep interva is .e0 micro-l t w ; ; ;
i i i:condn bu: thu prc';ressively decreases to O and l$ ( l i i
; ', l j ,
En ll.v 13 .it.s: before 2e earth is rea:hed. It would .! 7 3 !. ,! / 1. . ,! t e surmi:.ec. '%: the len~.t.hs ..are. c.orr.espondin.civ.
. . i.
V '
#.1 i, '. i " ,'
'ha-ter but. to :he sc iter's kaoveledge. goed evidence !
/ \
1 r N, ,
\.' -- - ' s ,
vf the strok: phenomenon near the earth is not .' N..,..' N / - c.vailc..o b. .:..ucpt ior the d!<cussion that follows - - r.
;mmedh:e:y, tF: st:p pher.cm:non will be neg!ected, po pe j j heaun it can have on!y a secundary ed::t upon the t, ,,,,u, m,z,,, , ,, , ,, i, 6:nctnenoa c:currin; near the :arth. Thus, as the 1:ader appro ches 6: carth, u wi!! be assumed to approact, at a censtant ra:e. c F c -
Howev-r.before having the discussion of steps, one ,
.s s, is i,
adJden:: coservatica shrid oe made. The steps, in the wrier . apir.mn. recuit from the centrasting . l
- ,! l : ,
I
' : i ;
ch ract:r of mo attributes ct. the leader., It was '
- l i
. . . , .' t, ,
3tated in ce: r.:: on with Fb;.1 ihat a corona space ,! , .
- nr.rge aeve!ce"- H :nivare ci the arras::d. highly '?
com.iue!:w c 'a.cri or core. I .-
, . .' .I
'n the . . r .:cr '. " Sea - imp"!se vehnte too I i l t'
Icv tn zatu.e ~ . .ver is .: pnp: t acmu a rod-plan: - l ! s .,
.s,.i
~
1 van, a char;ir; ::rr::.: 0 ::1. but w mr , . w :^f u on :, .
\
t... chercine :rrrec.t s a ..w -t-:ime imo. nica corot.a
. .r.e
,t .
.r,
. i i
ec car :ba n...s , cry rat h!!y to crest a:id then 1 i.
'. g' *f decrea:.es ;;cv - n z:ro. The decay cf $c corona ,
'- I' i, '
c rren: : M .aribi : d to tha d:velopm:,t c,t a caccaa n .:: :ncr.: tn ar. Ov :nhiti:s 6:rthr ib" r , r- 7- - ' ,, , r r - -- - -- r----->- oi current. >r md;;~. a w to o::ar h t.u case f:.e 2. .h.ts:* s m s.o c!. ,"r s ~ *W t ni. oi m: ;t a, e ca.n tut e;. ore the strok: current is w mbr.y ca'.d mf. tS: :,uce-;har;e a:movhere H av:e 4 arr <::si (or almost art:ste.11 g.r;; hic lly seashire e nd tr
- v o" :! c; . te
-: r na' iwr n s ':"orr.bie icr th: develorma,t (or c'rnnel current i., M ? p r N cr. .tes u '~r.-
c ner m c.,c:e ) i. th: :::annel. T'i< smte hu its sian u: "crv as: *.d E ning v ' andegy in t%: h.bcrt::ry rc,d-c!an gap when a va't:ge sl4ht;y in c.ic::ss er the critical sparking A?!9C AO NM TWi i!,',~.TH soitap is cpp.i d. Aft:r de mac char;e dareiop- F:g. 2 depkts the zu."es o' ^ :tmW m it . .1 rr.nt st:p chreit ccmaia:ed. Char.n:!s Ngin to the aarth i:: the 12 : s::re ae c '.mr.a < . . i 1 fccm. Tb cren aupp.;iaj :hes: chann:!s incre:ses beder is showa as e din u &:1 hre di:S -- , , po: nu.a. e ' . :aa . . .
. at : s,am; aM Nn fr.ter : an ever-in:r:ssing has .its up at h. .f,:.
i r::e. taken :s -50.01',03 vc!t :.% .' ' .' . . c.ctn- l In erh rep the lich:nin; ch nnel exter.ris in ward is al out one :.n: per mier mnnt!. ~' L i A > Q: ;o m c.: 8 a pencit < , chrma .nd :::oes v.h:n is shown apycaching : IN rr u. R" . 5'
.a
*7 t:p .s :0:6U I e PO;n: at 'Ll:a 9 . the space- cTie nursars are n .crie a .a tn :icus r*.
l
.hnr" c :Jnns ..r: ec l'ni:r conducice m the rJrouid be ren'lw + .. th: i..;e. *. car r:c r ar .
i fur *.h r 0:; . N.; th0 M r . "13. Ih ip Ce ch. rce "id:Iirni
.0 3 :0!::! 10:~ !!; Slov er 'd.:Ma and '$': f*C00is Fig. ':{ A ) ' 'n " . . . : ro, o .. : ,h: %-- -
r: mated. Th verap c::c' ct 1: en . r:e! cut- ;. i: r:ar' i ehh J . Op h - nC1t. . l
- nt mu~ 5 a : i cre'er cf 10 er 50 empire:.. the earn. h ir su"r:" m . ny . v .m c 1 m.
- , imth! vnu: is vo'ca'.'ry rv :h !:ss San Qis, tr.d sh:::$. V. Saa 0: .n! . sea wr.r: u..
C re ".: .:m cri a ir.:::ase er'er.a:b!'y. the :W , a .::: ma c.:c.=.: 0;(.r.s .P. .- i...,... - .,-. _.t, .s. . . G. a... i t.., N. .. ... ,
. . . . .7, ,~,.,..,..s.,
.. . . . ..w ! ,,a, .n.) 6. ,. . .
. t . , ,. .. a . . n. . .;..;,. ,
..s *
. g 1 e-g..,
$ .. -r4
, ;5 n.
w. t S o. g... ,,,, . . . ,.;..
,r,.
..;. . }
. s
., p s- .
i .- ,- t gy .g
,.>g.- - . - . -
o- o .
} .u .=> .
- 90004079 !
l
, . _ . - -. . . . - . _ ._ -. - -- - . . ~
i'
# . c. . .
laid down by the leader. And so 6. crc;;c" con-
,N s. l l m o6re ;
tinues throuch L Ei and finally ( _r ) . tne w : ant . e, g! N ' j "c- ' , jg which contact is made betw:en the ma;n char.d3 d ol s V \,[ Fig. 2(G) and (H) cl.cw : ter instants at wind *: C
.m , , ,
^
s. head of the return channci has nenetrated fanha- . e; into the space charge and lowered c furthe: ruct .-n j \gL,. ' of this char e to carth. The magnhude e' the cer-el C . i . M' n l. . A i i rent at the diRerem points is succened by t'e reis N 77 - I \ i i
. tive thickness cf the linec.
; .j j j ] i Note also the progress oi the t:pwrrJ them.e!. f.
a aicher 2na ne r-which ,s ..
- . i i i intended to succes: ~- - -
C.r - u u. ,S g,a ca o. veloc ty until the last step is compMt* estaNi*e ' i.O M* cfier which the vcbeity shou;d b2 cen<t;.n* er c:- ns. 3. I'k.wr ertwen irwency of occurrrrrte of licht. creas: as the head of the remrn strea rer sk:.vec , nir; > rrA s u!2 curve) <.,nl icioctry c/ remri; stroke toward the c'oud. At the ur.verd ci nnnel nr.';*c . :=. tentacle-like strearaers rea:n coward ard <iva*d and tend to spread the positive charg: cver a pre c:r arca than that encompassed by the c:n.nnd im
' 5'l
- CCEST cur.f!ENT AtG i
- v 1C'V2
' - -'u '
b nx wen u.o e:a i. bELCCITY OF ?.ET*JEN STiiCll2
% o.3 l , 7 inL-v in 'g 2
< : :.u; -
o KnowleJee concernine the m::'m.tua o f' h cr
- eu rent ci. the lichtniqu <t.c. c n o:. a1 ceur ~
;., - ~
& .. .:/ '
nature. Judemo, from the data ec.ta nigt se man f .' centa;: tnat exceed a per:ie"!:r "eme 't"r.m .; usuaiiv termed ticntn .. t.uon cru..- e.
- 6. .e ,m. . u- .
n ,a n. ca -
~ q - p *l at,p . H ess .
ek0 iS Quit C "rO!:-be.l?": I"rred '.;pOn Gr"t . p.r in n~ se lq 2 {Y.% ts ! .tue s ~:x!.e c... ren: crti . < .'ertry cf nU'nbCr Ol d;iTerent ytudit*% O d r the mo , hr r :: role c04ntrici But CO th .CthCT h0nd. th0 OtrCh; con' *i :, , D '.~.e m " D T.) (n"mber et *trm.o . .er. m.4 nem m: .. oo 6 . c :2 enly crudh hncvenelt is 'br. sed reca ihe n nd:: -
.er;. dim c::cuJs 6: cri:!:a! value of O'r~;t 30,000 ef rturm dau at the.-observen. lointi n; a cr'- : c4 .' ~
cb per e . do nor app;ct to be prrd:uiarly im- beine any day on vchich : cL.urv,.c b.s .....'.
]! i thunder durin; eny ei the inter n.! r! u hnh .:.; :cr- t pertant. P." .ee::ur on any prrj:ctic,n in the vicinity.
i~ r3e c3f t , ,,d :s 2:mer.t.J corona heath con. de:!s hh otbcr h..ud" routine oh cr rs..) .c T; 6. . . .. ... tmo: t?::r n: , :.6cna wa ard car.h. cad mean- tbc past f as" years wry prcrdr:n: ' e. ' .- .m v ,.de. tt. Mr n.. rpa,.e ehr.r;; abo'.e the ma t also undertaken, notaWy by 2. T Piccce.11 M. % : merras s. D. J. ' f alan anc D. '.hd: r-NiilehrnrJ.' in J : '. A :n;.:ai dat is r:nched in (m. at which the ve:opmcat cf an 1. strumr u tV vim cc at t. . :".m e'r ut stro;:es to ;; mund, in rce er tL u. v.preu x n 'meen h and the tcp of t.oe mat .m ~ - , nis cau . M.CO.0t') 6.000 = 3.?30 cm. or 270 crede measure in terms et sturm de n, tr u. teet. LaSya'ory te ts ,uh r,d-rod gaps having long has croved quite satufa< :ory in the r.:su .mt ! . t rpaciny icdictie tant t vw* :n the averap critical nin<: investiga:ca .1co.t forv.nra vth .;2.n v.; m 1 -- gradient of 6/100 '.cits per cm is exceeded. ch:.nnels to the d- c!opme: of a retisfactory ccuntr. cr. F:;ir e ; ra "cm beth noints. Th!! is the irstant in Fig. 3 v pb:ted the wre:n diar;bu:Wo cur ' . l
- a ' n::h " .h-avv rc: urn-troke current begins to develop:d b." AIEE ia 1M(4 I.cr cemide-t tr. , 7. . 1 forra,t.n' -
the referc~ : ncint for time civen at 21: hnown data a"ailable at th :!rr.c. in.: W - the tr,, c,. mp ,. rour mi:rosecond3 .,at?r. as shown of the n. turn stro,re 's of s a.it,e m ocerm h e "' r 1 in FO 2 C ' :h: ch;nnch frnm beth a and.h nave physical ph: nom.ma eccciat:J "t;. to re - l pregrmt m th; points indicated in 2tD). further stroke. In Fig 3 is rhown the dirt i nu. 3. . 1 procce s hac bt;en rana.: and the current has grown cecurrence curse of thH veMei:y er cWenn:t r -
~ l 8
i si gn:'i.:cr ly. At die same time the curt:nt feeding Wa;:ner ar.d McCann (Fie. 23 cf r;ference li) l tne down%:d moving ch nnd below h draws cur- a reunt of . heir analpis ci l'. a.:otrk pccun:,- - l rent frcm the c!J !cw-s.urrent channel r.bove b. S 55n!and. Mclen r.ad Cd!cm. '. rim or ..ary 1 TS: re.ucation ci this curreat is s:0,er than w::s unde:taken :er by ML ' Eibe"::.a. . tne speed cf hgm;it H limhd riimari:y l:y the pe:d if ve 9:c.pt the fact th't 'Scr.: : i. e s:< ' with . ' ::1 :.* etc - c.i :.m accommodaic itnif .a
. ic:eet rehnica .#xe:n ,t. d . :uire t -
)
I th: hip:- c',JacaQ .md ne sp::.c with ".hab it ve:caity uf *h return stret* ."' .t rc - ' cro ca $2 cher:e fru : tb ccrona sp.ce charge cbts;n dh reirt n tu?in : em:r :es .'i- ,
, . . = m m . =m . .
- m3 90004080 l l
i
-u c. nn*a? cf a:s ributien. The result is Ho' eser, during the trans; tion pa-isd :.0 ' -
m n S." ,' e * .hed curve et Fi;;. 4 ::: p th of the strc.he is accomn:ed.ai.:; i: ?!f . cm
.u u- ts 'ca- of what hap; ns in the re: urn the conductisity of a current of :bna: M ;-- ex
,re::: ., a e:enptes uom the earth can ber.t b: which is charae: eristic cf the chann:! e' the dw r-
~ cd a' y cennd::tne. ;irr: the s>"s::m 'of curr:nt . ward inder. to the conda:tFi:y of tre f ?.00', . m
" s nc. n m r 'g. .H ... . mese consist o our .
cres or so o the r: turn strod, e, a ver,/ hign irr.aven:
- r.:r. gular current waves LLat pre"encate with a o vohage drop must occur. The d am ur cf Q: . ire Ia-contani Yeiccay sc a!ang the paca;tel lines aa' and creases from about 2 mm to about 7 cm. Durine :S spaced r Mnce D. It is shov.n in references period the are must aSorb aa :no.~nou:. :mour.: 2
;2 ata 13 that :o esu6'ish and maintain a set of energy in in:rcasing this volume of ;as to the hi;n.
. :r:nts of thi.i ca r;:ter, an emi equal to temperature characteristic c,f at: cenduction. If (;.:
energy requirements derived from :e trauhng~caa
;' = w ,,', :n 3 - -
(2) relatisns are equa:ed to the enercy re+:iremenn @-
,. . e , ., .w.r*..' u' .n. ' . ' . ^ . '. *. . . o. . . .a...5.1 r 'd.- o f a a'. a tained from laboratorv t.:sts cn- c2 - ' s, e ir. Mb.:
. . .. . . <, 3.,s..,
- .. +. i .. .:,...' .,, c .. a, - ~.
t ,. , o t'.ed:,n ," e..'f to deve:c a rdation :wun the current ! t the re-
' t "- turn strcke and i:s veloci:v.R -
Thus. from the circuit condi.ica, it can be sen
', =
a from Fig. 5(B) that 6he power abscrbed m t'e bat-
-r in -
%:( t c , a (3) tery occurs ont./ during the perio ; c.:rir.; w.,.ich cur- 1
- -.nt dso b: inecrud at a radius a at a point that rent flows (which eccouc:s for th: ft.c:cr % in t:u e uation) and is ennsequen:ly
. ous pre"rchdy : cn; tne Im.:s 'n paase wi n . ..
': c. cads of Jr.;t -':- nt m es. This ' aves a
\
f, ,1 :, g , n , _1 ) 3 ./.' a e, p. r:maner: rer'n w.ha:n : qual to . 2 '
/ ;
' D The energy input thr.t c.ill brh ' an are :a i tv-
. = W.: b -
G) ficular d"rce of Oc8 T :* Ota be :P a mine * "Y
, . t
"* mDVN . ~&O
. of. In Curr:nt WC V:4 With th0 Plea:arcm.".! Oi(P 'O$ie r ' hi.' '-a,'
J* ; l' . r' . 3-
. .- u " .s .:.g y ec. .'; ...'.n y. t,... .. g.
. .s n.. t6. i .+.,a o
e3 ngg ega .":, n'3. Pr '"e-s
' *- * '1' "'
. , , , , . " ,..h,,,.,". .
~
.g, s .
. : r"l71a t' ' P. Q.,t:i .' e
, p ,,
Voit '.V: >C;cnie
. .O > . , .. ... ,
. 't
. M.L " *' i. . . .
.i.p
. .J.
- m. , r' L, d. u. '..,
f .C*'
' "*-i' " , ' s W" '"I
* '* "* *r" *t i ' * - '
..... .a..., u s l,. ., ) t ,, .' 'I*
. . . 66
.. . - o. ' .'- -.~'
r:ns htrDur/1 in --2 D .ve C t'h W! ;"'t *,; ' me:ntcu w!:S '.ce m th'.. tra":1 u. .n te,: velo::.:v ,"l. O J C :.
'7 li :nt:a. ral:L,,. '..I . !'.. D' s of . /
0.00 161'- d C '. '
- C1
,. ". ' ' . , J,. .*) .t. .cn e' .:. ,o,
,; .,3 . .o.b. .i o,,. ". r e r n , n i '**C ' ' "*' *- - *" ** ' C u *"* * "* *' *"" '\ O "*- -- *- *~~ +
"-**"~'"
- . . - .- . a. ..u. . .~. . &.
NCleate a AU 'On 'n'r: ate in currem *rui
.'. De !? n~ r:-?d S Cer::en*rS t VQ, at a .OCint ' hat .-.^^**
.1 c"u..:P
' ' - - '. " # .b
. r , o. .., i t u.s to.L. .. . e t r. ., C ., ....a . ,,, . a as. ..
re us. ,. .V.. ,C n * * ' - - m v o ' ' 'm" * *'- ' '. . " .
"*u**-~*7'~"-' .
"r
,,: ridn o nd i.n ,
tend c; th' " cave, the ,.oci-ta en as t .e ins:;.nt at wmch the a:: ;ivnneh c.p. aa .d.m. p s ::co. .c.w.it r:rtiens ipp.,y waca t.i.: to b: d:v:Inped. ,tne smt! eurreu ::.ar r t a '. . c tr ,'. wave ri:n ,a.narriy tr m o .mn a dopm.;
.c" 'n 9is ::.e '.h: mardiac vc! r.p ir spr:ad out A ' " '
. w It: entir: frc :t of the cerr:rt " cave, and for ..
Oe .une fina! eu r:n! nas th: rame overall ma,rni-a ..
'4
- ' '34
.i ud:. va 3
- -; .1.;
lf n us u en short at:; with currcats up to tens of b * ,. { :t c.. a ; w.tia,d< c ampere: .inc.cate that aft:r th: crest .. e
. _ + ... _
current is rer.:a: . d : drop :rm tP: arc d:::::.s:s a We@. - W "<+m- - - :- : mc q' tid.bc. In are':% .:n Cac, a microse:cnd. ~~ " " - -
. +
'a n, out 1.00' .c:tt act em. .r.c t'c. n rnor: grr.du-v.,..;,.. ,.,s.e
., .m. , e ., ., ,._ . _ ., .:.:;. c a e av a cau:n sm dar valu:s_. .ms sudden drop m, !.
."'.p is u, men : rated m. c g. b. ' mc..1 shows the v J:p acreu ..nd tS: c .rt:nt throuc.h a 9 foot red- : e, r v, r:d /.an. .' h.: n i m -" ed by a voltage sufficien: to C R., , y _ ._..,o p-m e "n
...8 . - . * .e . . . .
.6 f cs: b: ram; are shcv n in ' A ) md 'B' end ' eir a
c',l L W' L c r:;w:s 'n (C). T;'c c::ur r c f :N "ohage record i.: f j 6
- e : in;y se'Ilim to in iic..:: . drc, of 5') ;;v, wh!:S
/: ': Gond) :D 1 Jr'.C D aI c:. .'JJ Veh3 per cm. Mnd 3 j '
i C ;n"# T 0'..LLhh; ".O Cst r . !!'.a: : gr15:n' ,p.,,. .e . l ". .d .. . . . .,' . ' .a 't t' a ~
..'...t.I s
-'O: 1,.tur' ' .. J. verv c'::d 'y ':'nieu a v 'ue . .
(C1 tha: in: st e u' re:~' w.w .v
, r ~- .*'.'l . ' vol::. '~r ':1, cr le;s. i. e - c. .~ , / . :. -r , : o.
R O (h A y % gl . l
5 n>" pc :.n a;: ef us riiv.:tien. The result is Howeser, durin.g th: trans; don <:-icd u9 d e .
, W :' d c'.ed curuc 01 Fig. 4. are path of the strehe is accomt:cacap i:.a! fr. ,
~~"' ' d::n icf- of W :'tl happ:ns in the re: urn the cendue:hity of a current of :Fo i: ^ S a rc c.
" e . it e:cag:te .com de carth can best b: which is chanc:erisde of the chann:' < ' the 6.r-
" cm:d by -rend.; ring fhr :he v. stem of curr:nt ward I:' der, to th: conda :i city o: t; 0.0C0 ..
"es to..n m eg;. .n. .). T'.ese consist of four peras or so of th: return stroke, a "er; hi:ht tennde~.:
- .ur.;; dar terrea n es C.;t ,cropngate with a voka;e d cp must occur. The d:arnner J u are ii.
- c e::ct .e:ccay ' c .urg Be para;i:1 !!ces a.r' and creases frcm about 2 mm 'o about 7 en. Durire &h '
waeed a d:uence D. It is shown in references p riod :he are must ahorb an enormou:. :mout.: ci
; ' c.d 1.i th:* :o es:Wsh and maintain a set of energy in increasir:g t% volume cf ;as to :he ai;;;;-
. - rus of this ebra:::r an emi equal to temper:ture characterisde of are :enduction. If G.:
, g energ" requirem:nts denved from :'u tra\ clin;- can
;' = . H 'n -
- 0) relations are equated to the energy re.l.:! render.ts ,!...
. : c:iner.;d o. ?r, En: on ca::.n side or. o at a .
tained from !aborator" b:: cn cap '- is p G: -
.. . to deve:ep a rc,a::. en .i :wt:n tc,.e current m toe re- . .
w e g a u t m. .n.r r, ' < tr: s retard.:n; emt.
"' mm se and i*.s '.M s . i Thus, frora the circuit condi:!aa. it can be 'een ,
".- m ,t -
\ p from Fig. SG) thrit the rov.er abscrbed in the bet. '
r i ta - 0) . j u te:y occurs only during the pen.oa d,:n. .; ' .uco e r-i c rent flows (which cecouc:, for th: fLe*cr ' 2 in de M.~' .' J r *. i ~ .. .: raL.t:S :7 O! a pGins th3t eg' uag,jon, ) gg.d [c ggp.o;g; ,c n .g o
-I U (a !ir.: In ph05: V.i:h
- -' N - . 3
. J.. .. :r u m e: This !: aves a p, = 1 :., , , a [ . 1,) ,n [_ u ., ,. :
m s
..... . , 3,.g ,
2 \ ./ h Ut 'd3! o f*, ..Orin2 .4 O!J MJ T. Jn'- g Tde en0r?.V y e a
.. , .; - (;) ..-l s.. . is a r (....- .. . . n. .- . , . . . . . .. s. . ,.... L. ., . . - , . . . .
a . . . .
*~'"g,..
- . . . . - .. s.,
. n. :
r i
.- 9 WCv04 Mith *h: ,F. .
t;:0 e':!r n'. e gg.
.n m "'.',..%'.*'*ys'..-
3 i *
.n..,
.e . , - . . .. ... ..,, ,. . ., ' ' ~ ,
-.,,i. , -
. . . .. . as C.-t a, .,S . - so
. gi ' '- L. ..'.. ' u. .
I P1 ..! J m;*'. :"C. O ar.0 % r.hu *::tg - m
.; U.F'n' .d.N- "nd Vol( $t!: I, [c;Ord
- CM. ., O, i
.1 n e r:0,'r.s :;. .C.ip ac;ou h -
' . .g. .J....' r. ;0,'*r: j 2 b ' calues USCrI,'l .. f J., Ui. . .!r-w".u.. W, ' '"o, ..,,.
. ;" - ;:.,1 u7 o-. -q ( n ". 'r3"el Witjl @: */CIO:!s*1 ..
I *, H F*1 graf d. cr4 f. C",,*", ', r'.' 'a t : ' ' ' ~' - oi c.6 l0 w x *- -:c. * ,p- m
.......-........n.,..-.
. . . . . . . . . . sa s e. .5
, .. A. o e.,, .. ,,e e.
,/
*r cf . car is eb::ited. Curr:nt anc ve!:::. nse.' cer:- e a '. CCrN 0 ntr . tCS, mdi0al 3 4td ..On ." : 3f: m currta" mi .t o!. 0:r 3
". ' - : d a'. 3 OCint . n a..
ri d0er00s ? IM Vo*t0. 2. A J *
~-
.> . it. C " cf i.: :. ! D: W3v". uc. '.'
. OC R .
".O .R Cf suC A C' n": T. c
,..,,C. ...i
..A ,1. 3 c.., c. w, . . a. .m.,,. t3NQU Us I.r.0 jr,u .ra ,M m;h thy a:c ;n.annyg cyg, , ,.
i
. . a, e~.
.d atop u re.,..c 9:.ar re::nor.3 a. !.v when t.,.e to : Ove oped. ,:e .anr.M currcu r"rior ta ta, .
~ m.c: ". x. : ria 'b r:ly ir:m o with a slopin;
'n :hi : ..e :h: ::terdiac voi::e is ser:ad out - -
A m
' i. : art r: " c? che curr:nt. 'vare, and for , , A -
- L - '.-_ .
- . h. ai .ter.-m nas the rame overall macni- W 'C
- 11. .
e , a' b
% f .O. i. j,'/ I T- aa shm :r:3 with cu rents up to tens of : ct .c 2 5
- W. c: aur:et e: inck-'e dat after th crest 6
. . . .., ._ .. ,. .. .. .a .m . . . . rnp..t..,.,,, . . .. ci ,,. .... .a. 3 (..i,'. ./. . .,'. _' s. ,. ,'. p e , , . . . . . . ..
. . .m . .
m eac a microse:cnd. ~~~
.i. ' - t. :r em. .:.: then mor: grr du-m.e, a. . .:.
.:.m.
.._"*" .. ..z
,.nra c.c.,.:r "clu s. .,,.,,. ms su;. den urop in
. i, J;.r.m .r-':d . Fi; 6, hid shaws the ..
crm ; * ' u c . rcat trrouph a 9 foot red- ,
., ' n
.. n "" JSf a vo':ap mrT:ci:n to . C . , c . _ . 2, '.'-m"-"
y,..,. .
, r :..... ..cg , _.s n.. . . . i.,., i -
. ,.. .. 1
'AI -' a; ( 3 ) and ' l 7'. cir ' ' '
~ gr-"'p. . i r ', ' .......-.._..i.,..y.1,,..=.....,.f.j c f 3 .
. . . . . . . .. s... ....m y. g
. 7,y m 50::n ;o ir'i!: a: 1 te' et .~ 0 ;; , "c'#:S I, .
l'
. ..a %
e-..+..r. .**?
*J'
.I' \ .%..! .u- p r u* i. .
-..40 *
,/. . . . . ,
.\. -
.....i'g.i,... . . ..e.. s
.A s
gig s.. .g . I .4. 6 L
.lg h, 3.e q. +(.
g .
...,.e.
; , rt
, 4 .. f.,,.,
.f. ,
.. 4. -,..,.....,,,.._.3.
. . ... , ( ,, , g ,y 4 q,. .
n . .n , '-
': s. c- , :. r. v -
n ,m r--- - s- ,r 7, r D~ D D ' " 4 a
& f1 . .'W k
,m.~ ,
0 , . ,. . '
<. U- Uuuu u instant . re asumed :o be assc iated with the de- The foregcing nsrumptien-that the c"ene rc- ',.
ve'orm: : of the space charp. The current for this dias at which the char;;e m::h: 6: cened:::d c': - . .. case wa taken as 1.700 amperes; thus the energy centrated and the radius cf the cylinder ir. ,'.ich the ll be:om= equal to 0.0013 wett.tecend p:r em per current !!aws are identical-is err.'neces. , ce chey [ an rer: ?.verating the result: of a larg: number of is concentrated at a much larger mdius b cnd th ,'. n .,, .inin:r : L: en c , 6., and 9.foct rod-rod gaps gave current at ancther point, such as : ic.s c. cs c, :"a a re u : egaal tc 0.0020 watt-second per cm per in references 12 and 13, shen t!.i: Jn.r.e : n s .', amp-;re. . ac:uas rower input into the arc of ti.e made, equan.oa i becem s rett: n enk: nas: 5: proportional to hs velocity in ,..u, n, q ['[ trn per tec, which is r times c tirnes the current. i= - g- W Thuu 30 !n - 3 b
?, = tc:0.0020) i watts (6) ".;
In the in: ercs:s of 3impler presenta:!ca, cc ict;. i Equa'in; 3 and 6. going expression otr.s de',c!v;i:a tor wavce thr: :r: el y c'U C,, 9) t
. para!!el to each other, v.":r.? ,in n e ae:u ' . ,a -
i= ,n i _' t. . W v ave trave! uo. ",r.rd from a f.ai N. ane. I-io.' ' r h' ..
.w n - refer:nce 12 : k demon.::ated thnt an en;.rc -.
a
- dentical to equation 9 resu!:s for thisncase er . .. ',
Of regard:cl as the instantaneous dhtance frc n carth .t.
, that tac tip of th: return s:roic has trwe!:d. cr' o- - - - -
(8) Tc cristnin n tra!y:icri r ' den h::w. n 'nd .:< v m ...am s, . '
- c. the 'dW::in: o ra r.::c:s c i '. ' N-n~4 2 ,'.
n- - 1 -
.c a c
en n ci n : . 32-eta:c m re- :n , . .
. . e r- -
t. tN :n*:r o n.n. it a n.:t en , ' . . ~: .u e
. cur n it v. 3 D tu: l'c . ,e . : s . J 3 im --
- y. _ ,
~
uw- y , . ., . 3'..,,.........s..., a...., .c. . . .. . . o. ... .. ...,, , , , , .:. .r [ ...
- in:: t h :
- ro ? e e=c r *. rm.c c : I m .6 ri m . ' ie '.L r c ava. , ,c., o; .: Lc:ct , . .: .1 :: 'h : ' e r . '. , ,.
7 vcMn; c. tn: v2:: 3v,5 cf cc vr.,nc* cMeni and ",il F. J' 7, tnea os 0.1 m..u .r. r pumn: cyc: :..: ::. . 1 .
- e rm c: cf c=:nt:.. Tc: fr.::c . o is w." ' - . .. m
~ .~ . .u. met:. c~em.ov. m,my e a,n.e p u s: r:.r:"- c
. "_ __.._..-._-..__..__'.'..'_., i 1:en is avr.;1:.'::: cone: nin:. :n: dir:rit at::r. F * ' 5o - -
e
,n.
I ch cge areund rbe riou cr m; L ;..icr. T:- r "-~ c 7 . .
?~ dr.n ui2 b.: r.r d :hi: for a . - k: vtcre reu ;' V ; .- a
".m. .i .,. ,.. ,-; i,.
i i, cur en: is ul .timately equae to A.UJ, .a a i .. c:. v a! te mrJ: m2. :o w :n ud lv o. < o. . . :. . . .
! ! the radios wn! be prepactionr' :o the e;if n'! O ( '.n
'2 i I
i D, i Wi:h thes: wumptie is, th .w1r.e et.rs . c ' " - 1 is cb nined. W curve can b: com;'.ir;d -
-. .ac a ; por curve for the f:cm dc.ca.
Lund.oim." Fr k l y . a: J ict r NJi .-
] 4- ) ';;
P Nj , , brand derim' simi:r: 2xpro Juns 'x. ret. .::
'] , , ..
6
\ !]>l4t/f
\
empirical relatien cf ' Tog::rW .;ct -tn.u o: resisterce N tac p.:ta :: ~ trar cnec.ar: *
'I ;
; 0 ver:c:v r.s th,: cha:ge p; sir ' rc' .! ;r ".
- j 2
; **N I chr.rg$ m h. 'i hm c: t.!!r " cM en :r ..v-I
; equation S. :a ' teh o s.canant .' -:Mer ' c. ,
f 1
./ i ,Il l tity cia.03:0' /'0 ln (D,a). U. a; in. U = .
l
~/
3 l 9 (91.4 rne:: s) and. a. = . 0J f /': e .LM t , .
'2~ '
V 1 - hl , i
<: = 25 0,000. This em. pere- *t 900.000 sum::.' 1" Lun ibn: 5^n.. . F:
' ' . . ' .' :'nd 200.rC0 fer 9. or::: r.rcres.wr v I i
.! . .s_, . i . -
{ \f l the tr: urn - .o'sc. crd f 03.0' :n th : I i
.I, I
l c'! d,et c' rm:nt .n he . ":: .S mc"" r. i .'
" cre : ! he ' uMar. .*!!: ice a; c;; '-
2 3 4 0
- 2. r:nst:: : s . , ...
.. .. n.a .. ...... . ,, , . . ,,.
. . s, C , '. . ,*
, ,_ ; e.z.. * (,T' s 1(. . e ..;,y 'p ,., c ,4 ,
r.. . ..
', -t -
' . e? :* * * ;t
..? t. c. .! t s:r . .' " ' *
"., y??
p ,., , . . t' i %. m * * ' -
,1 m . .
= - . ,=
90004083 4 .e
i i l weferrn 7t de eart:nt of tbc return stroke at the l pund on. ,u d:te:ence et two negative exponential ,
;vrves. such u. ,;i~en by Druce and Golde" in which &
j I
%. curreni rbes rapidly at first. reaches a crest, and j '
I N.j u:n 6. cays siev !v ror purposes oi transmission- , . . l s i.ne compuedens this waveferm was usua!!:' ap- 50-
- l re::in:a;;d :'y a :inear!y risiar. g front to either 2 or 4 l
,,, ,' l nuctue:cnds v *:S a Sat top. I; was only after further c.n.e achton of the r:suhs obtained by Berger"in E
m bg/ N l
;m r-im :ateful, nOnt, and painstaking investi;ation a j op Mount 3rn S.avacore. tog:ther with an effort to 5 correlate these retuha with those obtained in the j'p,e E 3o. ;;.\ l .;;arr.tcry. : hat the wri:er and his associate. A. R.
1; l
- -Rhan.n, becam: cor.vinced that the waveshapes of ,.-de' E '
c i.','
.3 return dra':e currents were ccncave upward to
"': crea sa u: a-d then varied sem:what erratienity, g j. l.,,'/J
/
n:nJN; tren he ecm>,len;:ics of branchina. /
/e ? '/ l Fig. 7 is a repict of :he fronts of the larcar cur- - ,Y;'f /:% I rm re> cr:d M Derger" in !95' from his results w. ./ , .n hnt Lr S^hntore. Becai:se of the gradual rise ,,V< // / -, /'.,
a, ,/
^ c'c t(.:n i 2 'f t : t'roitt<. the actual ster!s wer A ; ,, A.. ,
- 1 U:im a. Tb r: fore, ihc curvcs were trans- e. !
t' co '. ' 'n 4 6 i.
'^
i
,:1 ('Mid shrot ;'t -;3.000 arr.p?r:s i the ' a t.: . . ' c t. iii f. S 5h0ws a later scr c3 of t,n e n m e t .s=:m fn,,. 7, ngor of ,,e a risgrc,.,, ~, 3 6,9 s ci.c in! q.::'ms b ;.r.ed by Dar;ct ' during 1C59 and rec?rded by Bcr,:n." .rm. :.c:::'
~,rr--:
,3 w t:. t me cr :o tv *
- p::ss thron@ 'a.'? A vnpc c: m v. m '.e & :. L Ci hnacc:. * = . a dis;urinn of this cues:Un i
is ev;.iuaFoa W i.: tur;; i~.redance of the stroke. j 1..
~: r=<ai--5 }' t..
downward .eaner in db:hardr.' into the nnal c.e wcer. e and of Fig. 2 oEer an impedan:: V/i(h i equal to unity. this exprmien :an be ured t, ',
' rate u " t: .. : dbehr. ;.: enn occur through ev:;uati' 're nup'r 'r. nom c ic , 0: R.e s.o v -
a .r.
.dch :he raurn stroi:: c'c ove :he fCISC'CO* d'0 I6EU7'"'2 enaN ' :"a cf c + d .
; o . ca . v,r ; ry. 2t F). can propagate within can be und to evalurt; tne impei.rce .? ..v.: . a J '
- ee c. ::a . ' n.m'. n .ny the downward lenJ:r. ba
- ta! 'n "- '00 "' ' ?
~t i . r' ped n. '
a mn; don ei th: h:i;ht of tnc tip i. c.mpe 3 fo w s a mu
'. r '""i p'
'.r i ' . 'ir.d J!>0 i!S v0!CCity. In refera g g.." t W
'D
..t?ct T' ,
l .,. . /. 0 IC')C '. ing af f rO W".i'!C 0XpfcS:h0M is . a .
. n f.x ta: emerdal ef the stroke: v. a numenc C.3 0. l
- Z, ohms 920 2 u'M
< i=
n nn o t c 15 33 ts
- 1!
g O_ 13 11 15 e t 9 ts :o - s atS
=208 T
-4C jl l *;q
*43j . .& .) y
.ta4
-60l -a:.
s -s y. .,,,. 6A l h5
.a . ,. .< ..
gg g a p ;3, y.! , , Csr fi*' M s :1 +1 ft
* '3 ** **
.. 3 9 'O '5 f* 25
% o % 11 '5 *S *, a 5 m "t '
r
~,,,,.,,--=s,, .aclN ~ - %. .,
2:l .zai- ~, .
-4C}
64 40( -
=rh
. N
'I Or72)dt3- ' .O
=f3i
'p-ha %3t124 s a?.172 O Du. !) 3' 21142-t 78 an. L/ ;* r .4 1 *:
, ., , m .
K , ,, ,, x _ S [
' ^ ---- s *29- ~*
**31 i
90004084
;I t':1/ J1 s.' t*'**:!*"'**
.rga ' vf Ut .
.g[ ' '
W (J c13' , '3 *t 't , W '
~ *I' l'~ , >
s.. D " W. ri;
- n -
P D A
~
g IL b d.
- 3. . -,, g;
- .. J
~ -.
l
The serie: impedance of the stroke is thus about current. and time is pletted fr.em the inrtant :.:r t I,300 chms. crest current is reached. It is shown in reference 20 that for rod-rod gapr- It will be ob<crvcd that r.li et th: :un er. h ..e &c betwear! 20 ; .ches i 30.." em ) and 100 inches (2M same ;;eneral cha;e in dat they are concave upe re. em) in 1:n;O. .h time Ing fer the application of a Furth:rmore. the time scales are t.t the ::ch* m: r 1.590-micrm:end wave is independent ct the gap of magnitude. It thus appears inut ;rp dice.mr;; '- icngth. T;u :ae gives seme justification for e.xtrapo- served in the laberatory can be appliec citee.. ' ta latic; fre 1 paps of laborater) dimensiens to those determine the nature or ;;;ntn:n.; phencment. The cf natura! lighmi: g dimensions. e.s ce!!ent :ctrela:!cn between the cherm:icm i.: r.e s l'ig. 9( ~d . e.ows the waveiront of the first com- laboratory a:iJ these on Merm Su Sal aar pm-ponent of current of lightnint; strokes. The letter vides confidence in the fact that ahhcuch the f art r desi; nations on the curves re :: to the corresponding were obtained by means cf a t 1 ciast tiop a hip' cunes of 2:37 The cunes in Fig. 9( A) are com- hill, they shou;d be chara:t.6 tic cf si:li!ct pn~ pated cu.ver beed upan: (1) an extension cf the nemena on flat 'errain as w !L E wot'c cf .1.kopian. Lariencv, and Torotiana in which
.ntv ;er.cn,treted that ter a 1,5-cm red-red gap.
cr o.r
" .~-m.- o "T S * "~"
- E ^- U- m *4 ~ -/' T" = f"m T-the telac:r, wig wi.ich the channe! tips from the To this point ccn d: ration hes heen p cn b:t: .
.v:0 elec.r<. des approach each other is proportional to the first component of a stroke. Of prmibi. :- .~
to the exces of the inst:.. ancous radient in the un- est to the transmi slon engineer nre 'he s*du.v.cr.' i:ndpd cap over the critictd sparking gr diem, which strokes. In Fig.10 nre >hown , nem' r er of odi~ u inderen.'n' ,f the e:m icacthi aad (2) the ns- gram- for >u'em e ent cern;, rant; 9 9 ob W n:c sumptw.n. r; , red H tests. that the c'.'rrent taken Ucrger." for :nc reriod ;9? . f er.r e c? b:. & ; r < crapert: erat to the veloei:" r.t which tripp'ng cf the e N;-aph. ir re- r,p - c 'e P -- the cha in:', arproach each ether. Fip. 9f C) and tain he entir: rmn fc- de Pr.7 cc' rw G ..
.De - the rnults ci actual tests. th. fctmer for by t !: i1 the fhst t to d'rir o ru ; ' :r : .
- ' the .tter fcr a rod-plane ;ap. In s.6 :c p:ent . alib etion of tha -H :ir:ri t . it - L :-
t o.'-r&i n. termined that the heat r.ust hr e r:er b d.c.n
.l. a...:. . rent " ciotted as a rc.tio of the f.nai
^ 0 i.
13 t i
/ 's Y
- r. g =," . ~; -
, ;e a -
... . rs e 2 a.-
-ine ts 'j < .
... m c3 "~~t -? 'V /
A* C$ h.i 0 S0 Y 'O6 ak.4g '-i. ' d $7 .--- ~~~ ., ' I J-'-
'e I-mc .3 2 c. e .'!a' < - - - t=
/r r9 6 C j , ,;- ,,
eW. , / ~'u .
/ !
/ \
t u.^ . *[, ' *,, ') Y'. , . , . . l l 2;CP,/ -02
- f. M' R
. . . - - . . . . ' ' ~ ~- t ,'_ ~_ 99 - *C l r - - - - - , n :-TAP '
.o ., 6 4
.. . ,: .: ., .s .. . o . .o .:
9 D
- t. - -
*'= -
l
*-I '
p, .
,f. r6 [*] %. e.1 e er - - .m y2. ,'., ..,.
j .a , .
+
(i E {
#r?
.b- -
r, r,,;x ,;rg
' " '~ ' 200 I C'c
.,
- 6 /: ' -
6C; 0 T C I'
- / '. ~ ,
f
-0. 2 ^ * " *
. ./ il. f 1
i .m . , , . ,ns 3 ---/,
-*C 2 3 /
-/ '** - 2,- AfI " " ' ~
C e w *$ -*. -4 -g e 13 . s im(. :n?p .m;w!an.c 'mm rde c. * #t r~. n r..rts - z ;' - n, V ~ -n c; r *!" #
.), ;c
< y,., -a.:e , rre. .v -ont. (. n " a.. i
-' v: .a no e er '. r t i,;l.. m ', * .9 ) l . ' *: : f.'. *1 I.' "C - t r' 1 f.h- a'
.~*- ' .: %- : r 31::" s g /~
p g~
~
- . m m q r' m '.rn'w -
- r. PN
_ . _ .4 - A .:3
' 90004085
~ . _ _ _ . . . . _ _ _ _ . . _
g, . . . r '
? ' l! g[ f
.-r
,'" Q 5 C. F. WAGNER -
Q
.- A
- c. =.3.,u..
- w .
n . v w w:..
.u- x The lightning strokey
}n@g.1 _. ..
as related to transmission-Ine performan'ce:'
% G;, '
- , L ..
PART 11
. 1 ;
R 4 .:'
.;i.
w.
- x. . .
a ,e . o u 4 , " - 1., l 4 pj h '. In the conclusion of this two-part article, the efficacy of ground wires ,
- )
i j, C in shielding the phase conductors is described in terms of the length of the
-;.,f striking distance of the downward leader. Upward concave return strokes, as well as f J f -
;17.1 repetitive strokes, have also been studied. The use of " pipe--pipe" gaps y-Y l across insulator strings effectively improves the lightning performance of the line .y T. 3! .
{ ',- -
- 2. -
- t. 1 ,
y y .
~.~ - 'i k4 ! ,
charge condition of the leader, computed the elec ca -
)
l V;.,19 i 1. It has long been accepted that the principal role of trostatic field strengths next to earth and assum' ell
~ l ffM 7; N :
' . the ground wire is to intercept the stroke. Then, with sutliciently low tower-footing resistances, it that the point to be struck is definitely chosen 'as j soon as upward streamers develop from the earth 2
$$ 7 has always been assumed that satisfae:ory transmis-or earthed objects. Strokes to earth were to be d y )
MN 1:# , sion-line performance would be attained. However, cisive when the computed gradient at the earth wa F i.. with the introduction of extra-high-voltage transmis- [ [M T ' ': sion and its higher towers, unexplained outages de- 10 kv per em. The writer followed another course and computed the critical striking distance from. V 'f Q veloped, resulting in the need for a re-evaluation computation of the stroke potential and the criti i
$f 1 -
5. of methods of computing line performance and for a inore detailed study of the stroke mechanism. , breakdown gradient as determined from extrapolat7 laboratory tests. Young, Clayton, and Hileman" a recent paper also adopted this approach in de
,: SHIELDING W..g f ,
^~
Provoost22 presented at Paris in 1960 an excellent mining the location of the ground wire. Their an
'.e sis will be discussed in more detail later. Eic 1 -
- resume of the role of the ground wire in shielding It would appear that the records of actual ex[eH h:D .h , '. the conductors. Supplementing this review the writer wishes to make the following additions to this rather ence on the operation of transmission lines woul rEr Q' o,-
provide the best measure of the protective angle'. 3[!D j 9',1- complex problem. is not always possible to determine precisely wh. g h/j :T.* models. Some of the earliest work was done on simulated It was recognized 'that precise similitude faults are caused by lightning and when this la C ,g p t is established it is not always possible to determ relations might not apply to reduced scale models.
,%-j.y The Writer, McCann, and MacLane,28 for example, whether the particular failure was caused by a di ,
),.;, Q"Q{.
stroke to a conductor or by a back flash. To suppo j*
':/.T in 1941 attempted to determine these relations by the conclusions reached with models, the writer, .
@NOhl;i, changing er the model scale anct by altering the rela-tive cloud height. The theories of stroke propagation 4 ;.n
,'?'" ' ';-
'W y'. 4" and spark breakdown that have been developed since ,8, 3,. that time would have provided reliable guidance on ,
p[ ' -
~
t Of 7'... the proper cloud height or striking distances to use. s c P O f.$ ' Another approach to this problem is geometrical in g. 3 7 4;c, nature, but its application is dependent upon setting ;
-2 F
i 'A' ? up a criterion to determine the location of the lower 8% . ..
.5
,l h ( ,: l T,b tip of the downward leader at the instant at which
%yn 7 "
. . gi.
ll % ;} @ the ultimate conductor,stroke termination-whether to the ground wire, or to the ground- tc the e . . . , ggy-li T '$c -* b,*, oc,fy ogegg, .f,. ?, ~
!'j-t.imji
- M ; .becomes decisive. ,Golde,9 assuming a particular
~
Fig. !!. Lightning outages as a function of pro $ctive n
; l &.Q * , Part n of a two-part nrucle prepared ortstnsny for for pubtlesuon 110 to 165 kv, vith average tower-footing resi.tta
, f; La certnan tn a current tssue at elektrotechassche zettscht,ts. - - less than 10 ohms, as reported by AIEE comminee yt jg- .
- c. F. wagner. Feuow rErz. is a consulting engineer. Pittsbursh. -
Wi .J.. y*i p i L).,:.,,;Q.- i E Pa.
,, ,s, e t ogt ,
90004086 '.e ,_
- . .; , b . . .
* . . ~
which might well be the case because of the low ~ . M. . ,-fg,' 1.., . 4 1. .; .'.__B . 3 .,. ' pl e -
&y%-g.l? '. ; "N
- ' tower'-footing resistances. * '.'
, c g.& l ,
55 - ~- - 1 - -
"' These conditions of shielding are quite consistent o . <.
l' 1
' ' with the picture of the stroke mechanism presented Y.3 7 ]
[7 . ] .- *
. earlier in this article. As ' mentioned previously, the .g. q p j '
. ' "C-f *~~~ ' ,
driving potential necessary to establish a rectangular . e. .;. a.h 1 . y( 1 7 :l*': i te 3 s.
"oovete.cmcuir " ' wave of~ charge and current that moves vertically ij
,5 y upward from the earth is given by equation 10, where f f 3, f' h j y is the distance from earth to the bead of the wave d '. ;Ib.c f 4, '
%g at any instant and b is the radius at which the charge '"'Tfg (!
? 2-Sucts-enw can be regarded as concentrated. Inserting the value Wi.f W Ff i.
s of i from equation 9 into this expression, then .. c.%y
~
M '
%o *
. **Ej ; f; l y
- u. .-
r n .o io zo ao ao so ao . } _ volts (12) N ?>h ! NEcME ANcLE V = 1.2 X 10s e_ l
' - Y' /
; {h Fig.12. Lightning faults as reported by W. Casson" 1 - r In *1/1n U b 2 - :@
ce f& ( D The parameter a varies linearly with current and b "
~
j %'l l gCann, and MacLane: almost linearly, if a be assumed equal to 0.1 foot . -j p . c' , jul 2 examined the data available and b 10 feet at 100,000 amperes, then equation 12 .'* df gl lh.1940 and published, among other curves, the one can be approximated by
.,Y@g k W.} '
. a j[down
; sperience of ina large Fig.11. number This of lines incurve the voltage was based upon the '.
'(13)
' .',:M:W.
M-<g' irange between 110 and 165 kv, for which the average v = t.2 x to' I ~ 2.2 e.- W .$.C ; e P r:
~ ' < tower footing resistance was less than 10 ohms. In awing the curve more weight was given to the This equation indicates that the stroke potential is , M:
4 khq dependent upon v only. According to the stroke points representing the larger circuit-years of experi. (ence. In 1959 Casson" presented the results of a mechanism theory presented here, the length of the final striking distance as the downward leader ap- f.q h g@ dh yh gd psurvey questionnaire, which has now been extended proaches the earth is equal to the stroke potential gand covers the performance cf 180 circuits compris. 2 ';, divided by the critical breakdown gradient, which is Ic'g 24,197 kilometers in the range above '!25 kv. 12 is taken from this report. The breakdown about 5,000 or 6,000 volts per em. The smaller the Q9 l p d,,.g5M final striking distance, the smaller must the shielding 4:~.h to' single circuit and d auble-circuit experience angle be to provide adequate shielding. Therefore, J. !!J T
'shows the efict of the higher towers of the latter.
to be conservative the value of 6,000 volts per em J. ~D M.
.N.In 1958 Burgsdorf: 8 s!iowed the same effect of is used and the final striking distance X, becomes
'.,D?@J - L. % %
h height of the line, but went further in that he %s% $ h..; geomputed the number of expected back flashes and, '
-x,= 2 x to' 1 - 2.2 e 2 em (14) f M
- 4. subtracting these from the actual flashovers, segre- - gg] g ggated the number of direct flashovers. This, of O F @}- C. g ., , .
gcourse, assumes an accuracy of back-flashover com-g-putation that may not be warranted 'in all cases.
= m t - 2.2 ,8 '"' II'}
' S'*[i$ Ih td./J, gKostenko, Polovoy, and Rosenfeld27 extended this This expression indicates that the last striking dis-
._h'% "
g [g ,g
+
pwork to extra-high-soltage systems and found the tance also is a function of v alone, or since v from , 2p , di' -*
.following empirical relation to express the quantity w g %'.}
%, which is defined as the ratio of strokes hitting A
I
- Q.@h ,S ;-- f f. ;
.the phase conductors to the total number of strokes hitting the line. oaa. Lec gg ,, - 4 . . ,yJ 9gg; .
a.
~
E . . Y ' a \/T -4 (II) cao- t i,,-41. . ~ , log 4 =
% $ .( r.h*
f'd," S? 5
' ,.Q
- 55$, -
*C .kCIC o.c o . n w g .w q W~ [ 0.o8- I 9.' y'.' I b,a=ground Protective angle at theeneters tower, degrees
' ' h)' .
p wire height at the tower, h/This function is plotted in Fig.13 for three differ.
.N bd..a
.k ..
gent heights. In the United States, experience indicates etbat the number of strokes to a line is about 100 per - --N.d WP :
" + :Q. h.%
100 miles per year. Taking this factor into consider-Q,ation it is found that the curve in Fig.11 agrees , - j g., i,f o >o so s 4o so 'e
@tremoy well with the curve in Fig.13 for 4 = 30 1 W*tters. The implication, of course, is that all of the pig. 13, p,os,ggi;,, of ,ng, gain, y,;;u,,, acco,s;n,
,o . {. &N go4,,nio, rofovyy, ons go,,nf,is" 'Fr tages for Fig.11 were caused by shiciding failures, m l
l Jtm E 1963
- EL ECTRICAL EhCIN E r. RING a 389
'fly i
[ 1! i 90004087 t w_ _ s
,sw,m w.f.;a s. a Wan,;s sd : _ -
= =
- S*4: SmwO"M**? **Y**N*-WYYY ,
;I - ,
. k .. -
r- WW ^f. t . l . . *
-J,.
P
- these SeC3L s.
- .e l theyc fi 5 e '
f ^
\
r, / /j D " -" I V3bes {,, , '- 4 u 0 . . carl!Cr
. 8 ' l [g 0 N
- h. ,, ". m ce~overca.I.." .2 7 / , I i em have I
-U MU l ~ ~-
- a --
~ ~ * ~~~S
! I I
l f0 4 ferenC
.' ~; \~~ t,
~ ~ ~-~l r, 'N - l 4 '_ for thi r
H ( ., - ( .r. l -
) A _
0 05
- l l n -
the ve l ',- M/M/////ro////M////Mf//////497/7MMM///MM///7M// ?* < C #Et Fig.14 (above). Assumption of stroke diversion acLording to Young, o o P w a 355U* i- *o L
~
Clayton, and Hileman* **CM ""M " DH3 , , pheno Fig.13 (right) Shielding failures according to Young. Clayton. and ward Hileman, with Mteen SM. inch 'nsulators; b = 13 feet O fg ':*y .
*y a /U .
~. .
Fig. 4 is a single-valued function of the return stroke .c., veloCII [ with the striking distance X, would, as shown current, X, is also a function of the current alone. param Walter, " attract all the strokes within a radius ~r ~o shown l Table II gives both the stroke potential and the the earth, so that r' striking distance as determined by these relations. M- failure The stroke potentials are of the order usually asso- ' - X. V24/x, fut ' 32 beight ciated with this quantity by other investigators in the 3 may b or within an area, so that _ sets of field. From the fact that the striking distance can ~~ . A - 2M - w, a scan fut
) securer
=
be of the same magnitude as tower heights, it can n - be surmised that the protective angle should be a back fi Thus a person 6 feet ts!!, if the striking distance
- The
.f function of the tower height. 300 feet, would attract all strokes within a radius t The direction of propagation of a leader can be
- two ot 60 feet.
j , viewed as being influenced by two diiTerent effects.
~{ $ sented If the assumption is maintained that the pr6jecyc y -
The first is the general directive field produced by countr. from the earth does not influence the stroke densi exposu the electrosta:ic field between the cloud and the then as the height is increased all the strokes wi earth, which determines the general path of propa. tion it
' a radius of X, should be collected by the projectiId attract-J,. gation of the leader from the cloud to the earth. It and a further increase should not increase the n '
,- second is atiected by the different cells of charge concentra. ber. This conclusion does not agree with experien. -
L' as dese
' tion within the cloud. In some cases this field will A height is finally attained at which the attra( than it.
produce cloud-to-cloud discharges rather than cloud. effect is increased either by simply altering the. ' This 1. I to-earth discharges. It provides the general guidance or by the production of secondary effects, such' of the discharge. The second etiect is the more ran. " value c
- increased glow discharge currents as premised to be c dom field at the head of each step. This field is Muller-Hillebrand.' '
f f - It w responsible for the sudden change in direction that For the moment let us continue to assume, . l tests oi sometimes occurs when a step is arrested and sub- Projections from the earth of the height of trans . sequently reinitiated. of shiel sion-line towers do-not- affect the un'iform str and no: u . ,, If the earth plane were. perfectly flat, one could density. The degree of shielding can then be. . e *
- of baci assume that the stroke density would be uuiform Puted on a purely geometric basis. ,:
4 k[ over the er. tire surface. In this case a projection from . In Fig.14 the strokes between B and B would 7
' i does nc The
[ t'
- the earth of height h feet that is small in comparison attracted to the ground wire, be: ween A and B ' the obs
[ , each side to the conductors, and between D an ; manipt Toble 11. Stroke' potentials and striking m cac s e to & cari Young n aP anad 4 @ ? Gat th distences of various currents . problem on this basis. They first assumed the re results g; ,v.. *"g' ' ' [ i. v. 1,, tion between v and i given in Fig. 4, computed -
** l stroke
- 9. , e
' **{"
a;
-] -
stroke potentialin accordance with equation 15,' transm u m took the statistical distribution of current into' ' i s.- a me 4u u m sideration. They determined the number of shiel 4 f a geon l
- l. I c.[.' .:
J.'*, ;, - E 2" U *S $ failures from operating experience, subtracting striking l 84 8 12 *St computed number of back flashes. and then 1 .
' l
1 1 i
~
. toese cata as a criterion of cecuracy of the method. .- . . . . . . , . -- : W 3
~T .
h -4
... ~2. The' length's 'of the last striking disiance 'v% @y
~
l!*they?
~
Because compensated the the results assumption didof not uniform agree stroke - with this criterion,. "..;with the return ary ? . ...j h
*dstribution and ^other assumptions by modifying smaller currents. . Of ? -
the ~-; -
-fQj %
1- ((airve.in. Fig. 4 to obtain satisfactory agreement be- J M./sh hp
- 3. The lengths of these,. strikirig distances 'a Ltween ' computations and experience. This appears the order of 50 to 150 meters..
'.M
.g
$sdfiable 'on the basis that the values for velocity 4. These lengths require that the protective angle liised in the production of this curve were average
*values over the entire return stroke path, whereas for a 50-meter to,wer be less than for a 30-meter tower. .
- 4
,j.T 'g (Erlier values in the life of the return strokeBACK shocid FLASHES 3- .; G
$ve been used. .
. C ,,].1 g r.rIMuller-Hillebrand' has shown a co'nsiderable dif ' Back flashes have traditionally been computed by
- f assuming a current that rises linearly to a flat-topped 'N.L ;
.s !
'.'t inence between the average velocity and the velocity Mythe first 5 microseconds of return travel. Further. crest value. Berger's waveshapes with their ever-
$ ore, a modiScation in one of the relations, such as, increasing rate pose new questions regarding the f.,kf g
rp]
~~
[ t$e velocity-. current curve, is justified because adequacy in.the of the linear-front assumption. In com-Tmputation of the potential, the return stroke was p uting the probability of flashover of the insulators, pp g , Ss'iumed to rise directly from the earth, whereas only the the front of the current wave and a few micro-
,5h M N j ph.enomenon is complicated by the simultaneous seconds up- after crest' current has been attained are of N
. .-'Q (
impoitance. However, to determine the destructive iward and downward approach of the beavy currents g 8 Nocity-currentYthe last step. Because curve were the rnost convenient action upon protective devices the entire wave must these modifications of the be considered. s
' D Wk lndg y
by [jitiameter to alter, Young et aFi obtained the results In the past only the effect of the current injected ~ M.5 i b; I ' on
!"fdifailures are plotted as a function into the tower has been taken into consideration of the computing ground flashovers. With wireshown the old assumption that in Fig.15, ;Min yV inhwhi i . j the current rises directly from a stricken tower with-W ; m)ght ~
at the tower and the protective angle. These }/.d ${L:jj ' be compared with the curves of Fig.13. Both out the presence of an upward channel, Wagner and g
> sets.of curves are, of course, based in part on data Hilemana showed that for a current front of two j?
7.~- P h ba'ck flashes,se. cured by assuming satisfactory ability to the microseconds, time-space change in charge in ' compute '5 # ts the stroke channel above the tower was sufficient 'M Md, of , WThe paper, based upon model tests,22 emphasized to produce a voltage across the insulators of the twar other points that in the light of the theory pre- same order as that produced by the current. But ff EAj C ( sented here may be of importance, r according to the theory presented in Fig. 2 the
,Q p@} hy ,
3" u in hilly ,3 V1b current develops by the formation of channel simul-y, ;y,cguntry, the down-slope side of the line oders caposure to lightning than does the same cortfigura-greater taneously from the lower tip of the leader and from ' [ il , in
)n 7 @rrlin level terrain. This results from the lesser the tower top. Fig.16 represents a simple model"
- e k t-M %' {Y .
v p .ttgaction of the earth to strokes on that side. And, of the current and charge fluctuations for use in T.h dM N estimating the effect that the charge might have on MN' e- gecond, in terrain of extremely high resistivity, such
'e ; OadDesert country, the protective angletheshould voltage he across lessthe insulator string. In (A) the ~1}5 d > Th{in terrain with low or moderate soil resistivity.
y p.g likewise results from the smaller protective - [; /.7
.s
; Elue of the earth, because in effect the towers appear,'
(, .- !):,d [c: Y _f@b;e
- w of greater height. l '
i- 5M p?* P I
, ould be highly desirable to make additional teStl on transmission lines wherein the single factor I
[.hd
' 55 ' f S
.LA i
's W! -
',- kh rh a(ndnot dependent upon supplementary computationsshielding
$)ack flashes. .Lj l
-I l ~'Y k@l;) .) .
s
- . ige w iter r does not wi h s
<j '
-H
' h ((. Y' b g #es'not exist, either with respect togeaofknowledto imply
'p De. observations of the strokes or to the subsequent
.p a precision that f ' '
' ,G.
,.e3
*7- F .g1 r
*5 -
Maaipulation of th i lT
~ N '
N.7i Ga e r parameters. He does believe ==^=" l
% l %
- 'h h.$
i I t, re,t,the mechanism of the stroke as presented here urts in a rationalization and coordination
~
M'
~ \, of the 5' ,
5 l l ,f I
- l MN
- 0 '
h.- om lI I i
.e.with the shielding failure observations on I .t i -4 n
; I. g sj} y
!nission lines. It has been indicated that I i
;! E@a (-
psThe observed phenomenon can be explained *on or W Ec .' h - ffd%% g
, netric basis through the medium of the last a, m f,n . , 8 dlstancC.
fig, M. Tirrie and place uriatiori of strMe current and
.[ M $:*
t charge just prior to and after impact with the carth , k-M
}{\
se ,
. 'i JUNE 1963
- y e.,
! ELECTRICAL ENCtN E E RING 391
$hi A,<. ~ .
,. .. - 90004089 l
, . ,.m . - ,.:...
r : j.(.a s7 M M s 2 5' M M 3 / N I f M n u k h.3 M. . . -.[?) $hD. g... .Q)dM. dM yV*"
~.. ,
- . - l Th.
y.
.r,e/ j <
~ ,
,g
' w
[ f
%',* 50 =
"~
drops in temperature. The density of the expanding <.A; the [. J. __ q_ cylinder is requires lower athan q disc. lower the ambient atmosphere j- a rarefied gas breakdown voltage- of tt
] g
,f@ql g t-0" j it may be argued that this is the discriminative factor
- th*"
ig~ Fe ,- But regardless of the explanation, the second anf 85 SI
,~ [ p. h $ l' To ._
cao subsequent strokes pass down the same path. It m'a* . rerit ) 4 . ;, - y sj curre ! I (A) superficially appear that, in a sense, the flashoverN inE D
]~;'.
path of apath of the stroke string of This to earth. insulators theory is notisstrict! merely an in th ex i UE _ _, e EU gg') gg true, hoivever, because in the early stages of. 't,4 that 1 i N discharge, the stroke current as it strikes the tra ' l the y pa Q88% a ggg ggg mission hne divides inversely as the surge impedan j is de.- {, ,, g gggg of the respective paths and only a small part ma}7 of sh j j g , g gg , 3 g i traverse the msulator string. And ,n the later stages; J s gap
'E
{ j [' AW-
-g ggg.g the portion of the current that flows ,m the conductor in seeking a path to earth impinges upon the ind The r % A a
4 .j
^
g (C) tance of the transformer, whereas the portion tha' y times. Fig.17, (A) Surge-generator circuh. (B) Voltage and (C) flows in the tower encounters only a very low lim 3vera! j current oscillograms for a gap having a len:th of 30 feet pedance. Thus, it does not follow that the.same cd-1 : , PVen l
- , y.:;:_ and spacing of 6 feet. with a series resistance of 973 ohms. dition prevails in the arc path across the flash -
OVCf V '
; ._' Scale: 369 or 345 amperes per divhion. 2 microseconds per insulator string as prevails in the stroke path. It ma a Part j
4, -j - di"Ui " be that during the interval between compone'rds- substa ( - are path across the insulator string has returned - ; Portion
.f .o is virl;in state. y. <f :icular '
4 i ,4- The effect of wind upon the stroke path of"s ,- portior f I. time variations of current, velocity, and the extension sequent components is to impart a parallel dispT" ment, which does not seem to affect the breakd
!! I0IIC 18 can L of the channels from the earth and from .5 in (B) ~
Y ,
..~ are shown. In (C) and (D) are shown the corre- characteristics. A corresponding displacement of n< PP OI h; channel of the are that takes place across the To a, 1 .- sponding special distributions at the instants marked s lator string would carry the terminals of the arc j of this :
( ors each curve. With these approximations it is pos-H f.- sible to compute a preliminary estimate of the effect away from the two ends of' the insulator string'a j mlatc (' i I influence profoundly the breakdown characterisu mitage of the time-space change in charge above the tower.
; of the stnng. ransmi g Oualitative considerations indicate that the effect is , . , .
One would expect for the foregoing reasons- p se th '4 d - not as great as with the previous assumption of stroke
,! Y. characteristics. the factors tending to produce repetitive insul j across current:
3 i ..~ flashovers are much less severe than those ten d I REPETITIVE STROKE 5 to produce repetitive strokes. And since the nu' cry that of insulator string flashovers on high-voltage.] ames tl l . It is still controversial as to whether a current of sig, i (Iso, fc [3 ., nificant magnitude, measured in several amperes or is only a small percentage of the total strokes to, d i$ -- tens of amperes, continues to flow in the path of the line,it would be expected that the repetitive num j
]1 . stroke be: ween components. Berger's' direct observa- . of flashovers is very small indeed , 'r; lare r ve er
~ ' ' ' - -
tions of current in the stroke at the earthis l'ndicate 900.40 o/- 'J 5,
.. PREDl5 CHARGE CtJRRENT3 A :hosen s
':- a value of less than one ampere in thein. terval be . -
Although a discussion of the predischarc,e eu - j 50 feet '4 d. ' tween components. This factor is of interest in shed-
, h; 1 F ding light on the reason for components subsequent of aps E is not directly related to an engineering
^
, available b:. $.- to the first seeka the same path. A continuous flow amination of the~ lightning stroke, no discussio
- Es 'I Ai 25 i the lightning performance of a transmission'lin tween of current may ns be a necessary' condition for such
,hi f.E. discrimination. Some authors have vaguelv stated complete without some reference to their influe [e aver 2; that the remanent ionization from the Erst compo- Predischarge currents are magnified in the case The d.
hh , e.cting th l>; #~C nent provides the guidance for the second compo . two electrodes that are physically parallel., W 1i- M nent, but have not stated precisely h'ow a greater , and Hileman28 bave applied the term '" pipe , appro: l a volt Ji M concentration of ions in the previous path provides gap" to such a gap, simply because their tests. ator
?'! $ the preference factor. McCann and Clark 82 have made with pipes as the electrodes and because
" udS-
' . . M ' suggested and given evidence that it is merely a term is descriptive. -
'4 Fig.1.7(A) shows such a gap having a len averag d.?-thermal phenomenon' Time is required for the very i]'i W.I~ ' hot gases comprising the are plasma of the return and spacing S, both dimensions in feet.' If an t q% am h' Mc. i stroke to diffuse into the surrounding cooler air..In voltage below its critical breakdukn value is ap 7 rrent rr
,iHe
! g this process the high-temperature core develops into to such a gap, then only a relatively small, danc
. i . . . .a i. . . .n a . . , . : . . $. . . a.. u.. no.. n.:, .. ... : n. .. .m ,.:,,,n, n. . .m om ator
.._...,,_..m discharge. However, if an impulse voltase in excess mam,$ ut tre grouna wire and earth. Since half q. n. - cr "
'of the critical valueis applied as shown in Fig.17(B), the current flows from each side, the effective surge' ?.C -
,'tben at some instant such as a a large current flows, impedance will be about 150 ohms. The drop through . 'M 2 (as shown by (C), until finally the short-circuit cut- this' surge impedance is then 2,700 X 150 volts or ~ .M. < ,
(rent of the, surge generator dows. The increase in ,400 kv. This drop is spread out longitudinally along fdI-current results in an increase drop across the " damp- the transmission circuit. In other words, to [Ing resistance" of the surge generator and is reflected a voltage of 860 kv across the insulator string, a produce . G.g ; , ]h in the sudden drop in the gap voltage. It is found driving voltage of 860 + 400 or 1,260 kv is actually .d:.5%
; fm / J that the average value of the current increases as ~" a -
T
'the " formative time" of the channels, a term which required: This voltage E,is plotte dashed line. An alternate method of comparison is to assume that the d' riving voltage E, remains equal >
is defined as the interval between a and the instant to 860 kv. Then, moving down along the E, curve to
)f )
&2 y .
, of short circuit of the gap, decreases. ' n' '
4 860 kv and then dropping vertically to the E, curve, { Fig.18 serves to illustrate the general relations for 'fJ,,y 4 . we find that the voltage across the insulator string: Vgit f 4 a gap 50 feet in length having a spacing of 3 feet.
$The numbers on the curve represent the formative and the gap is now 710 kv and the voltage can ap- flf pear for a period of 2.9 microseconds before flash , P, A e times. If a short-time impulse wave of a particular over occurs. 1*.$ .
,,J[T/*
Javerage voltage, but with duration of less than that 3df* MUD ngiven on the curve, is applied to the gap, then Bash- The inset of Fig. I shows the equivalent circuit, M fM over will not result. It has also been found that for which can be set up very easily in the laboratory. '.T M N' a particular formative time, the average voltage is hsubstantially independent of the gap length and pro-The improvement resulting from the
- pipe gap can be gauged by comparing the voltage E, use of the pipe .
- M
' }. ";A] 3, k 1 with the voltage E,.
pportional to the gap spacing. Moveover, for a par- . .I,k ' 3ticular formative time, th: average current is pro-
, In the foregoing illustration, a line insulation of -4
~
ci g,N i 625 kv was assumed. If, for example, a higher line J .g
. portional to the gap length and to the gap spacing.
insulation is involved, then the critical flashover pp, e,It follows from these relations that the curve of Fig. ~ . 91 %y "Q value of the gap should be increased correspondingly
.18 can be used to estimate the characteristics of a-the equivalent of mereasing the gap spacing. But .,~.'9 M j.* ;l i
. Agap of any proportions.
". To illustrate one application, consider how a gap the predischarge current for a specific formative time 2.
1S q0
~ 0h
[of this nature connected acro 3s a string of suspensionis proportional to the gap spacing. Since the surge C {h
; insulators of a transmission line can decrease the impedance is substantially independent of the insu- ['g., .d 'U lation level, the voltage drop through t.m surge im-
; voltage that appears across the insulators when the pedance is proportionally higher than the assumed g
. . /,ythat for a, ,giv fp/[
gtransmission-line tower is struck by lightning. Sup . .line insulation. It follows, therefore, ,
, pose that the computed value of voltage that appears length of pipe-pipe gap, the improvement is inde- g O'
'across the insulator string, neglecting predischarge pendent of the insulation level of the line. The ordi-
~J.Q'1 ! .rq p'
)
- currents, is the driving voltage E,. This is the quan.
?Q 19
,tity that is usually taket as the tower-top potential nate of Fig.18 can therefore be generalized to reflect this condition by adding the scale showing the ratio
]% 4@1%
(times the quantity one minus the coupling factor, ,45j fiso, for simplicity, let it be assumed that the waves of the average gap voltage to the critical flashover of the gap. To the extent that the insulation level (k{.y, , are rectangular. Let it also be assumed that the full- -;$ r_f jwave critical flashover value of the insulator string increases in proportion to the system voltage, it may .A M $g h.3, l I
,under discussion is 625 kv. This particular value is yhosen so that a pipe-pipe gap of spacing 3 feet and U$
,$ (p E
? ,jrd JO feet in length, for which experimental data are 7 20 j a.
ems 7' ' ,,, ' ',.. :QQ N y' availa ;e s l Inng.ble, can be assumed to be connected The solid line of Fig.18 gives the relation g,e across
, p oo.the
]_7Ed,p,,
, q
-Q q,jl [p.Eh, i * /
hetween the average current through the gap 0and 7 ex-
,, /.
X N ##
. . M, V phi
@c average poltage across the gap. I" j y- Q 20
# ~
*kh 3 t The dots indicate the formative time, which, neg-
{cting the time for the formation of the space charge,. g ac- \ ~ ,,}k '
$l as o gigsgsM3NS8c^br >:Q @kF' 's approximately equal to the time to breakdown.
f f'* - '
'Mj NN } a voltage of 860 kv appears across the gap and 5 a2 i #" - ~'
pulator string, breakdown will occur in 2 micro- ~l% @$$fi 7 ,.;
?conds. However, in the process of breaking down, 5> _., - Q fi.g 4
@ average predischarge or prebreakdown current of *o. o
. .JMu IW.n. L',
o D00 amperes must flow for this time interval. The doo em eo svenac,s .* exo cunawr 14eo 2ex n assess c. 7 n,
. j %o ,. -
hrtent must be drawn through some sort of II surge
' #' V""''""'"'
f$M
~
'h"'"'"' "I " #'#' * #'# ' "" #
, ?9edance of the ;ine. Actually, it flows through the having a length of 30 feet and ,1 spacing of 3 feet. Dashed O J L , jnductor and returns Wrough the parallel circuit line represents total voltage across the gap and a series 'Y,f] 'U] i resistance of /20 o/uns ,i ,, k J. f
- L JUNE 1963
- 1
. E t. ECT R ICAL E.N GIN E E Ris c .
393
; f.
% m. .,. p m =.m g + m y w w w w m ~004091 90
~
P~ ~
~ . .
. _ ~ r__
.r m - w @ @ i++ M M f h - d M " M N '. ~ ;.4'
\ - '
s.. U; w ' k
- .' &y. :
[ be concluded that for a given gap length the improve. ,. h$ u-<.. m
.f.h'W W i ment is independent of the system voltage. The question of furthering such a policy };'is4 K, !
Because the predischarge current is proportional the managements of electric utilities, who areEN1 b to the gap length, it follows that the improvement sible for the operation of the systems, and, to ;jIs should be proportionately greater or less than that extent, with the managements of the manufact ?- indicated in Fig.18. who must supply the equipment for the's How much of the burden should the res .> ' . ' J$ '
' An additional characteristic cf pipe-pipe gaps is of importance in their application. Tests have shown groups carry? What recognition should be 'gi - dF1 that their volt-ampere characteristic is unaRected by those particular manufacturers who spend tim I2 rain. effort on system problems in which they h{
g b~
- direct interest? These are some of the questiofs J l' '
. CONCLUSION must be answe ed by the electric utility nd' j u" 7:til-scale There is great danger that because of the' Transmission line insulation must satisfy three re- p it quirements. It must be cble to withstand profit squeeze some of these problems will ~ t - ' hat cover.
default. J; g
. ' and toaves
- 1. The power-frequency voltage and some degree .
- L w of Contamination. Up to now this has not been 8 REFERENCES 3 4 L [ffects of1 limiting condition and when and if it does become 2L Report on the Work of Study Committee o 3 (t; k h.mitmg. then doubtless additional research and de-and surses). A ppend2= IL The shielding r.rrect or _o, Earth Wire. P. G. Provost. Paper No. J14, CICRE. Parta.-
teso, O ystimated s selopment will be undertaken to provide a remedy. 22. shjelding of Transmission t.ines. C. T. Wagner 2
'astificatto; 2.Sw. itching surges. In some cases at present, Cann, G. L. MacLane. A1EE Traunctions. vol.
sis-2s. switching surges constitute the limiting condition. 60. g g,gt; 28 ** T'equency or occurrence and the oistr dution or t pd - This limitation can be removed either by eliminating ntng Flashes 1945. pp. 902-10.to Transmission Lines. R. H. oolde. Ibid / at-the high surges at their source or by controllinE them d A
- 4. shieidirg or Transmusion Lines. r s. Young. 3. M. cs '
l'or years it h by adequate lightning arresters. If the latter course A. R. HHeman. IEEE Paper 63-640. scheduled for publicat " ."
- is' pursued, then it must be remembered that the con.'IEEr ' Treuactwu. pt. tM (Potccr Apparasu and syste.g,a, -'f s
ificrease, swit- ;
- trol of switching surges can be attained by a relatively 2s Report on the work or study comrnittee so. e ik j #
Voltage formanceA-C or EHVTransmusion). Appenex. hird Report on une g hoe and aPPa few number of m.stallations. On the other hand, the 2:s xv. w. canon. Lines Designed to operate at Voltages arport uo m c:GRE.1982. switching stzrg provision for the power-frequency voltage or for 26. Lightning Prosection or overhead Transmtsalon lightning must be incorporated at every tower. ' recent Papers ns etcRr vra. tit. tesa. operating Expenence in the tlSSR.j V. V. Burgsdort. 5witeh' m g Sur
- 3. Voltages produced .by lightning. Although in 27. The Role or Ligntning strikes to the conductors s 1
the Ground Wires in the Protection or Illgh-Voltage Class. Distnbution C i some cases ljghtning does not appear to be a limiting M. v. Kosienko. L r. Potovoy. A. M. Rosenfeld. Elek .d condition, when the necessary improvements are have the resuh , made to withstand normal power-frequency voltage
."3 3** *; $;',[o[,',*'n $,j,;,,,,,,,,,. 3 M1 ~' UM" b schnft fue secuuche P%stk. vol.14. pp.181-26. de is to add I and switching surEes, lightninS will then become a ** ^ qf formance o Approach to uw caiculauen or the Lishenzng
"'"r Transmission Lines IU-A simplined M th ddi a ,,decisi informed , i problem. A line with a phenomenally good operating e o tr g , i to Tower c.
III (r'meer r. Wagner. Apparatu sadA. R. Mneman. systems p . vol.AIT.E
- 79. oct.1 Traunctases,$,y j performance from all causes should be regarded as sas-4os. - _ : De Usulation ]
.ng-surge duty. t an overbuilt line. As imbrovements are Eto$ressivel E made in system design, it becomes more and more "F.Line.
ston Wagner, C. in " Gas Dtscharges and the30 "' Lisht I whether ; ade g strokelewi supply industry" tdook>. autterworth .. mance anest imperative that the detailed reasons for the perform- scientine Puducensenej-I"f,"[,'** Recovery e charsetensuca or targekd te economic ance be better understood. And so the electrical in. Mecann. 3. J. c' Art. AIEE Traunction. vol. 62. Jan. , - Although res l dustry is faced with the need for greater effort to 3 collect information concerning lightning and to Nert or Predischarge currence upon t.ine Perfor $ Wagner. A. R. HUeman. AIEE Paper 62 1181, scheduled for understand its erTects. ucanon in IEEE 7 ausenou, pt. III (Power Apparatu f as shown that
- cemsi, von s2. issa. high-speed rech ersted by a rest.
.s; M
,_ IIST METHOD
-e . e Test Environme.
V l} because (1) the PLAN TO ATTEND THE 4 r der laboratory
.- ":' ind (2) the eqt
.~
IEEE SUMMER GENERAL MEETING w " m,. .ce. outdoor
' :..U b Use ROYAL YORK HOTEL v %nuauy ruu text mtata and oi
'ttE Technical opei
._.s trt w TORONTO, CANADA
~$
'Ne **" inter
. Us"l.** ' ' ce JUNE 16-21 90004M2x '
' for~med much o f-
'W.
7"
- b=*
. -* " Kaib. Fellow I
-,.,,H. ' n. ohao.
[ ; . I
. N* 41 E11GT n4 23
/
Kq y - i
.i COSIMUNICATIONS PAPERS !
t l !) i PAllAMkTHES DES COUPS DE FOUDHE PAllAMETERS OF LIGIITNING FLASIIES i par K. flcucEn,11.D. AspEnsox by K. IlENGEH, R.ll. AsuEnsox
'l ct 11. KnGsiscEn and H. Knussscr.n du Comiti d' Etudes n 33 of Study Committee No. 33 (Surtensions et coordination de l'isolement) (Overvoltages and Insulation Co-ordination)
Rapport publid d la demande Paper published at the request du President du Comits of the Chairman of the Committee j 31. V. PAI.VA 3Ir. Y. pal.VA i
- 1. - Introduction. 1. - Introduction.
l.c Contiti suisse des rechcrches en haute tension Ever since 1943 the Swiss high-voltage research exploite depuis 1943 une station de mesure des phs. committee has been maintaining a lightning measu. nomsnes de foudre au 3 font San Salvatore prss de rement station on the JIonte San Salvatore near Lugano, Suisse. 1.'objet principal de ces mesures Lugano, Smit:erland. The main purpose of these . 6 tait l'en regist reine n t des formes de courant des measurements was to record current shapes of . coups de foudre qui frappent deux pylancs de ' sis. lightning flashes striking two television lowers on I vision install 6s sur cette montagne. Des photogra. the mountain. During the night photographs were
- phics de nuit ont its prises de coups de foudre se taken of lightning flashes occurring in the vicinity, produisant dans Ic voisinage et l'on a Egalement pro. and more recently recordings of the electric field cide plus ricemment h des enregistrements du champ during a thunderstorm were also made. Reports on p electrique au cours des orages. Des rapports d6eri. the measuring installation and the results obtained vant l' installation de mesure et rendant compte des mere published first by Berger in 1955 (1, 2], and I r6sultats obtenus'ont st6 publiss en premier lieu later by Berger and Vogelsanger in 1965 [3] and i ;
par tierger en 1955 (1, 2], puis par Berger et Vogel- again in 1966 (4]. In 1967 Berger contributed one i sanger cn 19G5 [3] et b nouveau en 196G [4). En of the five articles on lightning for a special edition 1967 Ilerger a prdsents l'un des cinq articles sur la of the Franklin Institute Journal [5]. foudre parus dans une sdition sp6ciale du Journal du Franklin Institute [5].
- Finalement, les risultats des mesures effectu&cs Finally, the results of measurements belmeen 1963 entre 19G3 et 1971 ont sts publies par Ucrger en and 1971 were published by Berger in 1972 (6) and i 1972 (G) et l'annse suivante (7). Au cours de pan. in the following year (7).1)nring 1972 some of the nie 1972, une purtle des donnees recueillies pendant data collected over the period 1963-1971 mere ana.
la periode 19G31971 a its analysse plus en dstail; lyred in more detailt this time a computer mas used un ordinateur a eth utilis& cette fois pour d&termi* to determine cumulative distributions for positive ner les courbes de fr&quences cumul&es, s& par & ment strokes, and for negative first aud subsequent strokes pour les coups de foudre positiis, pour les premisres separately, and to calculate relevant correlations. composantes et pour les composant:s suivantes des Representative lightning current shapes and a sum. coups de foudre n6gatifs, ninsi que pour calculer les mary of all results are given in this report. I
, corrstations correspondantes, l.c prJsent rapport pr6 sente des fortues de courant de ftiudre et donne un rEsum6 de tous les r6sultats obtenus. -
- 2. - Parametres des courants de foudre. 2. - Parameters of lightning currents.
2.l. - !)ifinition des ratigories. 2.1. - Definitions of categories. Un a trouve que l' ensemble des coups de faudec !! has been found that the ensemble of lightning frappant les pylanes du llont San Suhatore pournit finsbes striking the lumers i valore t
) , . !, ni Ki.xcTa4 N' il 1 , &tre group & en ca1&gories distinctes, comme il est may be grouped into distinct categories as shown I t indispis ci-dessous, beIom : i Coups de foudre W"'"E 0"'^ #8 4 '
l I I upward pashes doscn unnli pashes coups a.cendants coups desandants , , I I I I i l l coups 1 coups i coups l coups positive negative positive negative positifs negatifs pontifs negatifs paslu s pashes Alu s palm - Dans le pr&sent rapport, la convention adoptee The convention adopted in this report is that the l, est que la polarils d'un coup de foudre est celle de polarity of flashes is taken from the polarity of the i la charge du nuage et elle peut facilement itre iden. charge in the cloud and is easily identified by the lifise d'apths la polarits des courants mesurss. I.es polarity of the measured currents. ('pmard (Insbes, coups de foudre ascendants, qui constituent la ma- which constitute the majority of the flashes at San jurit6 des coups au Stont San Salvatore, peuvent ilre Salvatore, nmy be identified by the upmard bran. p identifies par le fait qu'ils se ramifient vers le haut, ching of the flash if a photograph of the flash was (luand une photographie du coup a 6ts prise, ou, taken or, failing this, by the continuing currents of
- sinon, par la pr6sence de courants de quelques pen. a /cm hundred amps lasting tens or hundreds of taines d'amphres se maintenant pendant des dizaines milliseconds that may be folfomed immediately or l ou centaines de millisecondes et pouvant &tre sui- after short current interruptions by one or several vis, soit imm6diatement, soit aprhs une br6ve inter- impulse entrents. Domnmard flashes, on the other ruption, d'une ou plusieurs impulsions de courant. hand, branch domnwned and do not produce pre- l D' autre part, les coups de foudre descendants se ra- discharge currents lusting more than a fem milli- s mifient vers le bas et n'engendrent pas de courants seconds. '
de pr6dscharge de durse supsrieure a quelques mil- i lisecondes. Etant donn6 qu'on pense que les coups ascen. Since upward flashes are thought to be primarily dants sont essentiellement associss a l' effet des py. associated with the effect of the television towers , l6nes de tslivision installes sur le 5tont San Salva- on 3fonte San Salvatore, the analysis presented in 1 tore, l' analyse prisentse dans le rapport actuel traite this report deals e.rclusively with domnmord flashes j uniquement des coups de foudre descendants, con- which are believed to be more representative of j j sidsr6s connue plus reprisentatifs de la foudre na- natural I/ghtning. In this category, however, a flash I l turelle. Toutefois, dans cette catsgorie, un coup de may consist of only one current stroke or a succes-y foudre peut ne comporter qu'une seule impulsion sion of such strokes separated by intervals during de courant, ou, se composer au contraire d'une suc- which little, if any, current flows. The latter is
., cession d' impulsions sspartes par des intervalles de termed a multiple stroke flash and in tuch a flash temps au cours desquels il ne s'&coule que peu ou the first stroke displays characteristics which dif-pas de courant. On dit alors qu'il s'agit d'un coup fer markedly from those of the following strokes, l de foudre multiple et, dans ce cas, la premi re and thus when determining the characteristic para-impulsion pr6sente des caractsristiques diff6 rant no- meters of lightning strokes. these should be separated tablement de celles des impukions suivantes, de into at least Imo separate categories, i.e.,
sorte 'que lorsqu'on ddtermine les param6tres carac- (a) first strokes, and i tsristiques des impukions composan* les coups de (b) following strokes. foudre il est n6cessaire de repartir eles-ci en deux cat 6gories au moins, u savoir : , (a) premihre composante; (b) composantes suivantes. A la condition qu'il n'y ait pas d' autre facteur Provided there are no / nether variable factors variable exer ant une influence significative sur le ereccising a significant influence on the process of processus de la decharge de fomtre, on devrait trou. the lightninu discharge, there should nom, within ver, dans chacune de ces cat &uories de courants de these categories of lightning flash currents, be a coup de foudre, une population de courants homo. homogeneous population of currents, and all mea. gene et l'on devrait pouvoir s'attendre S ce que l'en. surements could be expected to yield uniform re-semble des mesuren fournisse des r&sultats unifor. suits apart from minor differences attributable to I mes, en dehors de diff&rences mineures attrihuables residunt errors. 7'hese results should therefore fol. i des erreurs rssiduelles. Ces rssultats devraient donc lom some distribution (e.g. normut or lognormal) ! pr6senter une r6partstion donnse inormule ou logo- about the population mean.
\
normale par exemple) autour de la moyenne de la population. 2.2. - Choli des paramitres. 2.2. - Selection of parameters. I.a forme des coups de foudre et de leurs compo. The lightning flash and stroke entrent shape may santes peut stre caract&risie par un petit nombre de he churncteri:ed by a fem parameters, which are y < 90004094
N' 'll stECTna 25 param&tres, qui pr&sentent &galement de l' int & ret utso of interest when considering thI' possible da-lorsqu'on conshlbre les dommages suserplibles d*&lre maginu effert of a lightning stroke. 7'he paramrfers caus6s par un coup de foudre. I.es parambtres me- which have been measured are the followinu . sur&s sont les suivants : , (a) Coup de fondre. (a) The lightning flash. (i) courant de cr&te - pic de courant le plus &le- (i) peak current - the highest current peak in v& du coup de foudre; the flash; (ii) dur&c da coup de foudre - dur&c pendant lii) flash duration - the length of time during laquelle il y a ecoulernent de courant ou, dans le cas which there is current //om or, in the case of a d'un coup de foudre multiple, le temps qui s'&coule multiple stroke flush until the completion of the jusqu'4 la fin de la derni&re impulsion; last stroke;
, (iii) intervalles sans couran t - intervallcs de tili) no. current intervals - intervals belmeen ! temps pendant lesquels il ne s'&coule aucun courant strokes of a flash during which no detectable rut-mesurable; rent is flaming; (iv) charge du coup de foudre - charge totale tiv) (lash charge - the total charge transferred transport &c par un coup de foudre. by a flash. ' # (b) Composantes d'un coup de foudre. (b) Lightning strokes. '
I (i) courant de cr&te - pic dc courant le plus Elev& (i) peak cur. rent - the highest current peak in ' d'une impulsion; a stroke; !' (ii) dur&c de front - intervalle de temps compris (ii) front duration - the time interval beImeen entre le point 2 kA du front et le premier pic; l the 2 kA point on the front and the first peak; i (iii) dur&c de l' impulsion - intervalle de temps (iii) stroke duration - the time interval belmeen compris entre le point 2 kA du front et le point the 2 kA ' point on the front and the point on the de la queue on l' amplitude du courant est tomh&c tail where the entrent amplitude has fallen to 30 % - 4 50 % d'e sa valeur de cr&te; of its peak value; (iv) vitesse de mont&c maximale (raideur de front (iv) ma.rimum rate of rise (current steepness) - du courant) - tangente A plus forte pente du front the steepest tangent on the front of a stroke; d'une impulsion; (v) charge du choc - charge ele &trique transpor. (v) impulse charge - the elettric charge trans-the par la partic variable de l' impulsion (le point ported by the rdpidly changing part of the stroke ; de la queue de l'onde de courant 4 partir duquel la (the point on the current fail after which the charge charge transport &c n'a plur et& consid&ree comme transported is no longer cortsidered part of the faisant partie de la a charge du choc > a ste d&ter.
- impulse charge
- was determined by inspection of .
min & en es.uminant la forme du courant et il n* cst the shape and is therefore not precisely defined); donc pas d&fini de facon pr&cisc); (vi) charge de l' impulsion - charge totale de (vi) stroke charge -lhe total charge in the stroke, Pimpulsion, c'est A-dire la charge du choc major 0e i.e. the impulse charge together with any charge ' de toute charge transport &c par le courant conti- transported by continuing current after the impulse; nuant a s'&couler apr&s la fin du choc; (vii) energie pr& sum &c de l' impulsion - Jnergie (vii) prospective stroke energy - the energy qui aurait st& dissip&c par le courant de l' impulsion which would be dissipated by the stroke current s'&coulant h truvers une r&sistance d'un ohm, c'est-h- r flowing through a one-ohm resistor, i.e. fi'dt Ats or dire fi2dt Ah ou 3lohm. Cette d&llnition fera l'objet J/ohnt. This definition is subject to further discus-d'une discussion ult &rieure pour accord. sion and agreement. 2.3. - Ripartition des variales. 2.3. - Distribution of variates. La repartition des mesures individuelles (variates) The distribution of individual measurements (va-des param&tres dans leurs categories respectives rintes) of the parameter;s in their respective catego-joue un rSle important de guide dans l' estimation ries is an important guide as to the vaInes e.rpected des valeurs h esemupter pour les mesures futures, to be found in future nwasurements it may be infer. Celles ci peuvent &tre d&duites, soit d'une connais- red,,cither from a foreknomledge of the physical sance pr&alable du processus physique engendrant process leading to the lightning discharge or from un coup, de foudre, soit la r& partition m&me des the sample distribution Hself. Consideration of the
&chantillons recueillis.1.a prise en considerution du physical . process, as far as me understand it, does processus physique, dans la mesure on nous compre- not provide compelling reasons to assume any par.
nons bien ce dernier, ne fournit pas de rulsons tienlar distribution. The sample data itself must Ilie-conduinant imp &rativement h admettre telle on telle refore provide the clue to which of the known dis-90004095 s
l 1 20 E O4 Tim P ll I t I r& partition particutibre. Ce sont donc les donn&es ellevm&mes prises comme &chantillons qui doivent tributions run be fitted to the data with a sn//icient deurce of confidence. Historically the logarithmic servir is faire d&couvrir celle des r& partitions connues distribution has bre.stne generally nevepted and nsst qui peut s' adapter n elles avec un degr& de confiance I annignes li! showed that cunm/ative distributions suttisant. Dans le pan &. 'la r& partition logarilhmique on a logarithmic base produce a reasonable fit of a st& u&ntralennent accepthe et des unaly ses ant &- the data to a straight line. The togarithmic distribn-rieures [O ont montr& que la courbe des fr&quences . tion has therefore also been assumed in this analttsis. eumulics sur une base logaritlunique s'a p pa re n te Cantion shontd. however, be c1rreised when using suffisanunent avec les ilonn&es dont on dispose, n the implications of .this, sinec there may we# be un troe& rectiligne. La r& partition logarithminne a other distributions which fit the data just as well i donc &l& admise dans la pr&sente analy se. Il fas or better. I toutefois prendre garde, dans les d& ductions qui en , r&sultent. au fait qu'il peut tr&s bien y avoir d'autres r& partitions con venant - aussi bien ou m&me mieux ] n ecs donnees. D""b *b~3'Y i OJ &
- 2. l. - Risultats. 2A. _ ltesults.
s' Les figures 1610 &tablics avec une base logarith- Figures I to 10 shons, on a logarithmic base, the t' unique, montrent les courbes des fr&quences cumu- enmulative distributions of ten of the lightning cut-1&es de dix des param&tres du courant de foudre. , I.a li;;ne droite a &th tracee en nainimisant la somme rent parumeters. The straight line is the Icast squares fit to the points (which have also been con-des carr&s des distances des points (&galement reli&s nected by straiuht lines > and therefore may be used entre eux par des lignes droites) et elle peut donc to obtain the best estimate of the mean p and stan.
&tre utilisce pour obtenir la meilleure estimation possible de la moyenne et de l'& cart type. e des dard deviation c of the logarithms of,the variales in the population distribution, logaritlunes des variates b la distribution de la ,
population. , Cela implique que la foncti,a de densits de proba-bilit& w(x) ait la forme sui. ante The probability density function w(x) is therefore i implied to be ;
#c (r) = (1/x o,[1 n) exp ((log x y)2f3g 2) (g) et qu'elle soit donc compl&tement d&tinie par les and is /nlly defined by the parumeters and c.
deus paranihtres p et c. l.orsqu'on, compare les valeurs r&ctles aux valeurs, donn&cs par la droite dt r&gression, il est 6vident ll'hrn comparing acInal values with values pre-dicted by the regression line. it is clear that sub-ipic des ditr&rences appr&ciables peuvent apparaitre. stantial di//erences may be [onnd. In some cases, 1)*ns certams cas. en particulier lorsqu'il s'agit des particularly where the e.rtremes of the distributions twtin estr&mes de la r& partition, on pourrait &ga- are concerned, the actual values might also be con. h ownl prendre en consid& ration les valeurs r&elles sidered when predirling freynencies of orrurrence Imor la pr& letermination des fr&quences d'oc c u r- of the variab/c. t i nre de la va ria bic. Il nii pas 6th trac & de courbe de r& partition pour No distribution of total negative flash prospective fenerme totale pr& sum &e du coup de foudre, stunt d'nne que l'&nergie des composantes cons &cutives energy has been drawn since the following stroke
' h premibre sont insignitlantes par rapport n celle energies are insignificant when added to those of , .t the first stroke. The results are also given in Table 1 1
- premibre. Les r&sultats sunt &,:alement pr&sen. I showing the values for the enmniative distribu.
b' dans le tableau !. qui indique les valeurs corres. (" nd.mt . respectivement aux fr&quences cumul&cs lions at the 95 %, JO % and J '7e probability levels
' " ' *tes uneaus de probabiliti de 95 %. M 'i respectively. Lts read off against the regression
l' ' line).
tres valeurs & tant lues sur la droite de o .re unin t Ln ' c 'pd concerne le tableau L les vuleurs des
~ mh de crite de la foudre sont corrobor&cs par Regarding Table 1, corroboration of the values of
' peak lightning entrents are shown in the distriba-
- dnf rnhutions de fr&quence publi&es par popo-
% . qui a trouv& une valeur m& diane de tion reported by Popolansky (81 who observes a
' ' V 'd tenue ;u partir de G2l valeurs de mesures median value of 23 k.t obtained from G2) values of reliable current measurements mode on tall chim.
'n' ot t ti ddes etivetuces sur des chemin &es et neys and Imvers, including 192 values from Herger.
""' dc ;:rande huuteur, y compris 192 s aleurs
'"'d This result tends to anguest that the measured paru.
dn essais de K. Herger. Ce resultat meite meters of dmunward flashes strikmq the Iclevision
, , ' 'ine In paruw&tres mesures sur des coups lowers on the mountmn in I.nyano may be compa-
'"A r 'b scendants atteignant les pylbnes de t&le-
"de 1.8 ruble with those of lightning striking tall situctures montagne voisine de 1.ugano peuvent in open country.
. 90004096
- Me 41 xi m:TiiA 27 '
T Ast. eat! I - TAHl.E I - Paramitres typiques de la foudre. Typical lightning parameters. Pourcentage de cas dipassant la valeur indi I,,%. 8 N Parametre Units IYr cent of u6e dans le tableaucases exceeding tabu h,o. Parameter Unit value
'95 % 50 % 5%
Courant de erf te Pent; rurrent (minimum 2 kA) 101 Premi6tes composantes 1 kA 14 30 80 negatives et coups de foudre n6gatifs . negalise first *trokes and fashes composantes suivantes l 1 135 nigatives kA 4,6 12 30 e negalite folioscing strokes couis de foudre positiT3
} (sans e mp santes suivantes)
"6
~ kA 4,6 35 250 positive flashes (no follou' ing strokes)
Charce 2 93 P'"'*#'## * " P **"I'8 C 1,1 5,2 24 negatives negative first strokes composantes suivantes 2 122 n6gatises C 0,2 1,4 11
, negalist folloscing strokes 3 94 * "P" de foudre n6gatifs C 1,3 7,5 40 n egain.te pashes 3 26 C UPs de foudre positifs positwe pashes C 20 80 350 Charge du choe impulse charee 4 90 premisres composantes C 1,1 4,5 20 n6gatives negating prst strokes composantes suivantes 4 117 L6gatises C 0,22 0,95 4,0 negatites folioscing strokes coups de foudre positifs 4 25 (""" 2,0 posita.seule nt flashes impulsion)
C 16 150 (only one stroke) Dorfe de front Front duratwn 5 89 premibres composantes us 1,8 5.5 18 n6gatives negality prst strokes composantes suivantes . 5 118 n6;atises ya 0,22 1.1 4,5 negatiw folioacing strokes 5 19 coups de foudre positifs ys 3,5 22 200 posntire pashes 90004097
;. g su:ctnA N* 41 N* 4 1
*/o TAnixAu I (suite) - TAnix ! (continued) i i 99 -l i
' ^
Pourcenta::e de cas dJpwant la valeur indiciuse dans le talileau 95 Vig. y 1%ramitre Unite
.%. Per cent of cases exceedine tabulated l Ihrnmeter. l' nit rajue s l 80 !
95 G 50 % 5 "r sli/dt madmal ~ Alanamnm duldt 6 92 50 prenneres composantes kA/gs 5,5 12 32 - n6galises negative prst strokes 20 -, composantes suivantes negatises et coups de 1 6 122 i foudre nigatifs kA/gs 12 40 120 5 ' negative following strokes - , 1 and pashes . '
* "P? de fedre positifs l l 6 21 kA/ys 0,20 2,4 positsee fashes 32 -
Dur6e de l' impulsion 5( 5trole duratton , 7 90 premibres composantc4 ps 30 75 000 ; i n6gatises j negalite prst strokes l, coinposantes suivantes l 7 115 nsga:ives ps 6,5 32 140 : negatire following strokes. c"Ps de foudre positifs .' 7 16 ' posattre fashes us 25 '230 2 000 l
; \
Intferale (;2dt) lettecral 02 dt) '
'[
premieres composantes 9 91 negatives et coups de A2s 6,0 x 102 5,5 x 104 ' 5,5 x 108 foudre nfgatifs negative first strokes and pashes composantes suivantes 9 88 nsgatises A2s 5,5 x 102 6,0 x 102 5,2 x 104 negative following s,trokes 9 26 e ups de foudre positifs positive pashes A2s 2,5 x 104 6,5 x 105 1.5 x 10 / Intervalle de temps entre 99 - 10 133 imPul 7ons negatives ms 7 33 150 Time unterrrds between 95 3 negatste strokes
' Durse du coup de foudre Fla*% durntrmt gg .
nigatifs (y compris les ' 8 94 coups de foudre i une ms 0,15 ~ 13 1 100 seule impulsion) 50 - negative (including single ~ stroke pashes)
- nAgatifs (non compris les 20 -
roup4 de foudre a une 11 39 wule impulsion) ms 31 180 900 negatine (excluding 5 ' single stroke pashes) .
~
11 24 ", rna 14
, 85 500 w
IC 90004098
l N* 11 st.m:Tair 20
*/o I j\l Fig, 1 U"'I l
~
I i Courant de err,te - l l
~
i!( l (1) I'remidres impulsens nEratises 4
,' .i j (2) Impuluuns manantes n ratives ; ..i, a (3) lanpulsions puestives ; } l ll? "
t l Peak current I (1) Negatuer first strokes t (2) Negative follosreng strokes
- y Bo (3) Positi,,e strokes 50
\
\
bkl\ 20 2\ i 1% , T 5 D D D h
}
l t, f~ - 1 I ioO lo' 10 2 kA U!p i \i ' P illi9
!! l@
!8 li Fig. 2
? !1! l
\ \I ! I 't!U ! !_ !I lii, !!!!
ig i ,, ; i 'lin
" \ lily \ IIP i lllll! :!I
* \l ll X l ll l !lllll ll Figure 2 9 % Hh i Nvil! I
. M \%31 I inh l I Charge de l' impulsion - Q (impulsion) 50 :lih2 Y t h; !II iMI !
(1) Premieres impulsions negalises g 3 , ;, 4 +tf (2) Impulsions niivantes nigatives (3) Impulsions posibwes i g ig.l j4 l
, ; ; j Stroke charge - Q (stroke) to :
l (1) Neptive first strokes l' (2) Negatis'c folloscing strokes l ~*\ lI, l ~'* l (3) Positive strokes 0 5
.i I+ . . - bE' '
Ib ' fit. Y sY ' ! ! f . Il 1 ' lid I l \lil l l l l UIl 4 Fig. 3 - io ' so' io' io 2 c l\il ! , i s >
% fr% i li so 1 \ l l \
1 i h H 50 h \ l
^
$ ---g - ;' g, -
t yp - "*""
~;1
+ , g p Charge du coup de foudre - Q (coup de foudick
\
j j j (1) Coups de foudre nhsatifs, I ; 7- j -(3) Coups de foudre puntifs lj T[l li flash charge - 0 (florhJ (Il Neptkr flashes e go*' 0 t l0 lO' to C (3) Positwe J1*shes 90004099
3 11 Et.miliA N* Il s - 11
% . ijli p Maure 4
- i i Ya "- ,! [- l1 j 99 - t li l ; 71@* 4 ,, Charge du choc - Q (rhor) Q + ;.t j p- - Tj (1) Prenurrn impul.aong nigalises 99 - l b} :
*t';
-L m impula una nui antes nbgativ as _
95 - _,_ 1-
+!'~.' h **-
h t..t) Impuldone pu4tises 93 . ~ ~ lw - +R' 'l' l- -l l l l Imsmlse charge Q (impulse) (I) kgatu e first stinkes lf I l l l l (2) Neestive following strokes . 80 7 1l { (3) Positwe strokes go - l]
~
j l 50 t.
; p, V~ --
i i I p._ ; j _.
.4 50
- g. - --
20 h:i .-
-+
2 IN-l 4 l 20 -
'lll . h r.
E 5 ! s __ 1 m -'- tt, _
' [ T' i . -
i - L h ~!' i ! g
~
.n tPA i n [
10 0 j - 10 10' 10 8 C I I I! l Fig. 5 to 4 x !
! I UN
__h . Y N-+-.-Wt ' lD kN' Id
- Figure 5 80
\l, "h l l J lY-lt Durie de front - T (front) Ti 3 3 (1) Premieres impulsions nfgatives (2) impulnons sunantes n gauves lN , ' (1) (
(3) Impulsions positives, 50 I
\ (O '
I l \l _ f ront duration - T (front) (1) Negutive first strokes I \ Ii \ .. - (2) Negatwe following strokes Eo l f _ lh .Q (3) Positive stroken } ) l l
}$\I .
5 l ml A[
-ct:
1 1 1 ii vimvi i1 I l o, 1 Ii.16 i I i IIi\ i 't i
! ! l l!!H M N!! Fig. 6 io-t to o ici 99 ' ! ! I !!!! !N J Al ps -
! ! ! U!!i M\NT 99 -
93 [ ;
\
' l ii!! 95 Am% i _
80 i ii 80 --
. ... N. h .
d--- i'{! t w i ' W% y!! s c 5o
' ~
. .lb i3 iRy '
N a - Figure 6
~ + 1,.3 - - $ ~
- I i l 'ill l ll '(-ZZ lMl j Viteuc de montic maximale du courant - di/dt (raideur de ~
*17U
+w a.*S wa .
front du courant) (1) l'rrmiires impulsions negatives $ _
'{ T
, r
}it --i ;
~** (2) Impulsions misantes nEgattvet
, * *.,i . . . .i ,
l l l
. 1, (3) iminilwons poMlivet
....,,., Il !. I ._
Alaximum rate of rise of current dildt (current sterpnosol - { \ l \ l 'l (l) kaatwo first stroko n
'O,, (2) Nowntive folloiting strokes I
(00 10' 1 - kA/ps (3)'Posaive strokes Io \ 90004100
I
) Es.rcTHA 31 f N .61 f
?
'/o Figure 7
' I Fig. 7 y Durie de l'impul*n - T (queur) 99 g ( l (1) l'rrmirren innpul n,n* sdr. liers (2) Impul+m, e nte. rreativea y
y g 3
, _i (3) Impulwn. po.itnra g3 ,
i I l l'l l Stroke duration - T fsad) (I) .Vreatwe firer strukn l'y l f:) Neentwe falinwine strokes 60 g 3; y ,, ,;,, ,,,,,,, 3 mr x 5
.; \ N\
] l l
20 S 5 1 1 1
'I ! 1i !
g I l \b ! I i . i i !. i i ! i. I L iI ; I , ilil
. 10 10' 10 8
ps !W ! 'lhd llM ! l'HM I! Fig. 8 ) pm !. po! ii. ,i !.uiq l!,ppy , , . , , , ,
. !ll I !HTl I ! !' "! 1 ltilR ! l ilihl ! Ijh!
l 4_14,,A i 1:. i 1ii. . 3; . ! . i !, .a t ! ! i m: l lII l ' ll1113 O!! l l Ih l h! Figure 8 80 lb . i Durie du coup de foudre - T (coup de foudre) l!li ll11l Nllij j IN l h p, (1) Coups de foudre nigatifs y compris les coups de foudre gg gj gi. , ; ;y g gj ; jjp (2) Coups de foudre nigati e on compris les coup's de foudre 50 sarnples. !!D.
!44 b\
I . T 1I I"I ! ' 1 l (3) Coups de foudre positifs. !il . !I k ! P'l I T IU! I h I' lash duration l'(flash) gg lblI hIU b l
,} b N (1) Negative flashes including sinete strokes I
,1 l._ yI e (2) Negative flaslacs excluding single strokes ,,.
2 (3) Posstice flasines lllll lll,f
'" fidl lll,l l{
5 Tk ih ii i h:i i u ;j i %_u,, llH1 94 ;l;n l Illis it f, O I ho W he; ihm pc lI" illi lin I .h.I !An l IIii l IIII 10-8 0 10' 3 i pn 10~' 10 10: 10 ms
" __t IU , \ dthlh lid. I I In ii [H lU:
tw3 % ti! Hi !! . E, 95 d' I i lk [ I l l0 lld h 80 lh l h; k h ,i *--
' Eh l$
h d lIn 1 ll!'i. I l ll4 If iH M ii ll I Y I- Y !.. l[ Di' l$ il liU llig jT lll$ 1 im ' f 20 b l
'k E Figure 9 f III 'h i l$! ,. li: Energie prisumie -f i ds
]i
**~P (1) Prenuires imput. ions eigatines 5
;p
" -l - : * - C ) I"' P"I*'""* * " *'"" O *' " "
ap. . . . " . . .i g.~. (3)- Impulens puminn
~"fiq ' Hip i1,. a, O, 7 l y} ,
"iiey4._
ii jp l i li:, IYo'Pertive enew - f i2dt I t h 1 & {-TX
~
T~~1T" iIl .%5*'ive fi
- t'""*
l 1. il l ylO l 11 ! liiM i !iiH
!!hi ik ' # # ^ *" U" I"""'"# '""
(3) I'osntire stroken 8 8 8 10 10 10* 10' 10' 10' A8 90004101 l
4 1
*l3 n r.ctnA N* 11 ,
s
/* Si l ! I! Fig. 10 m N l i i i ' :
99 4
-Ags .. ,
r k
! l 1 l\ l i .
80 : 1 \ l !
! X ' i i \ i i 50 , ,g ; ; qq q g. -
! \ i ! i D q
.dJ.. m j %p j ! eo o -
20 l 5 _ u . 1 . 11 I ( j i Figure 10
!! 1 P !
intervalles sans courant entre impulsions nigatives N IU \ ' l heurrent intervals between negative strokee 0 8 2 10 4 ms 10 10 10 Otre comparables k ceux des coups de foudre attel-gnant des structures de grande hauteur en plaine. ,
- 3. - Corr 41ations entre les pararneties du courant 3. - Correlation between lightning current para.
de foudre. meters. ',; 1 i.. . On pourrait concevoir que des proprietis impor. Important properties of the lightning discharge could conceivably be brought to light if some of the ' ^~- tantes de la dscharge de foudre puissent stre mises en lumihre si certains des parambtres cit &s au para- parameteri mentioned in the previous sections were graplie prseddent staient suffisamment corr 616s. Si, significantly correlated. If, for instance, the peak par exethple, il existait une correlation lin&uire entre current and front duration were linearly correlated. g le courant de cr&te et la durie de front, on pour. a simple time constant to determine the rate of rise rait soup onner l' existence d'une simple constante might be suspected. Simple regression analysis of de temps, qui permettrait de calculer la vitesse de peak current and other parameters, taken one at u mont&c. Une simple analyse de r&gression conside- time has been carried out before [li} and the results ,,Q rant le courant de cr&te et les autres parametres, have been presented in the form of sentter diagtnms k puis s&parement l'un apr6s l' autre, a 4th ctTectuie showing also the linear regression line. Linear re. ( autrefois (6) et ses r6sultats ont sts prssentss sous gression of the logarithms of the variales was assu- f la forme de diagrammes de dispersion comportant med, i.e. the relationship y = Aza. Such a pictorial , I mssi le trach de la droite de r&uression. On a sup- view of the regression is reassuring in the initial poss une r6gression linsaire des logarithmes des sinues of an analysis but for the purpose of merely qN, variates, ce qui revient h dire qu'on a admis la testing the hypothesis of correlation between ' the relation y = .tz". Une repr&sentation graphique de ec genre de r&gression est rassurante dans les pre-parameters, or transformed parameters, it is suffi, cient to compute the correlation coefficient and test { mier's stades d'une a n alyse, mais, lorsqu'il s'agit its siunificance at a predetermined level. seulement de faire l'essai d'une hypothsse de corrs-lation entre les paramhtres ou entre les parambtres transformss il suffit de calculer le coefficient de corr 61ation et de v4 rifler su validits k un niveau pr6dstermind. l.cs tuhleaux 11,111 et IV donnent les matrices des Tables 11, fit and lY show the computed matrices ' coetilcients de corr &lation calcul&s, respectivement of correintion coefficients in respect of positive stro. pour les impulsions positives, pour les premibres kest neuulive first and neuolive foIIoming strokes. impulsions n&gatives et pour les impulsions sul. vantes n&gatives. {' l.es tubicaux de coefflcients de corr &lation sont The tables of correlation coefficients are consider. . consid6r65, co.ume pouvant apporter une aide utile ed to be narful as an uid to the formulation of n ) h la formulation d'un modble de d& charge de foudre, model on the lightning discharge. Some pnrnmeters, 90004102
.c 41 Es.mTna 33 T Anu.AU II - TAHLE 11 Coefficients de corrilation entre les pararnitres des premiires irnpulsions pmitises (c'est-d-dire de coups de ;
foudre po<itifs. car d n 'y a pas de coups de foudre pontifs multiples). ; Correlation coefficients between parametern of penitisc first atroken. (or po itise fla hes, there bring no , positive multiple stroke flashes) g h ' r Vitesse Durie Courant Durc,e de charge du de Charge Charge de Duree de du i de mon tr.e de choc Energie l' impulsion l'irnpul4on C
"P coup de
I ""'
I
'"*** ?'
Impulse Energy 5troke Stroke foudre curre I d $t n < charge charge durahan I,4,", a c , ofrise duration e t Courant de crete i l'eak current I'00 l Dur6e de front I-Front duration O'O ' I'00 Vitesse de mont6e max. 0,49 - 0,68 1,00 Max. mte of nse
*'F * *" 0,77 0,23 1,00 Impulse charge 0.27
" *' EI " 0,84 0,22 0,39 0,82 1,00 Energv Charge de l' impulsion 0,62 0,32 '0,11 0,74 0,72 1,00 Stroke charge Durse de l' impulsion 0,58 0,48 ' 0,02 0,80 0,72 0,75 1,00 Stroke duratmn Durse du coup de j
foudre 0,33 0,112 - 0,15 0,24 0,17 0,64 0,24 1,00 ' nash duration Charge du coup de foudre 0,62 0,32 0.10 0,74 0,71 1,00 0,75 0,64 1,00 Hash charge ;. I Les coefficients en caract res gras dipa*sent la valcui csluque au niveau de meni6 cation de 5 % Coefficerau e~n buidface exceed she ecstwel sour et the $ % level of surmileance. Iberves de hberte .11 Dessers of freedoen . I1 Certains des parambtres, tels que la charge de choc sneh as impulse charge and total stroke charge. are et la charge totale d'une impulsion, sont li6s l'un related as a resnIt of their definition and a good . b l' autre en vertu m&me de leur d&linition et il est everelation between thent is self-evident. Neverthe-dvident qu'il existe entre eux une bonne correlation. less other correlations, or the lack of sneh, may pro- l N(anmoins. l' existence ou l'a bsen ce de certaines vide significtml pointers to the probalde discharge autres corr 61ations peut fournir des indices signi- mechanisnr. ficatifs quant au m0canisme probable de la d6 charge
- 4. - Formes typiques du courant de foudre. 4. - Typical lightning current shapes.
4.1. - Giniralitis. 4.1. - General. 1.a constatation que les formes des premibres The observatiun that positive first stroke, negative impukions positives. des premibres impukions n&ga. first stroke and negative following stroke shapes are lives et sles impulsions suivantes n&uatives wnt distinctly different, particularly in so far as the nettement ditV&rentes. en ce qui concerne natum. front durations are ennecined, has led to the desi-ment les dur&cs de front. nous n enuduits b d&llnir ynation of three categories of strokes. Good cor'e-trois catigories d'impuhions. l.a bonne corrilation lation between parantelers such as peak current, I i 90004103
34 stucinA N' 41 TAutIAu til - TAat2111 Coefficients de corritation entre les pammetres des premiires impulsions nigatives. Correlation coefficients between parameters of negative first strokes. Vitene Durie Courant Durc,e de cha m Charge Charge de du de de mon te.e Durie de du de choc Energie l' impulsion l'imput.4 ion c uP coup de [')y' f",",', ~,,' clmpulse I:ncrgy Stroke Stroke ';, , foudre current duration rate charge charge duratwn gy,,g lash
.ofrise duration
' "'E' Courant de crate Peak current '00 Durse de front 0,37 1,00 Front duratwn Vitese de montfe max. 0,36 - 0,21 1,00
.llax. rate of rise Charge de choc 0,77 0,25 0,39 1,00 Impulse charge IE i*
Energy 0,88 0,30 0,42 0,89 1,00 Charge de l' impulsion 0,61 0,29 0,19 0,91 0,78 1,00 Stroke charge Durse de l' impulsion ! i - 0,56 0,33 0,10 0,51 0,52 0,43 Stroke duratton 1,00 Durse du coup de ' foudre 0,08 0,25 - 0,02 0,14 0,10 0,23 i 0,11 1,00 Flash duration { Charge du coup de - foudre 0,54 0,36 0,19 0,72 0,64 0,78 0,44 0,64 1,00 Flash charge i les coefficients en caracteres gras dApesunt la valeur critique au ruveau de signification de 5 % ) coefficients he boldface escred nho crawal value at the S % level of agnificenes Degres de hbertu : 77 Desroes of freedom 71 observse entre les parametres tels que le courant , impulse charge and fait duration confirm that sha. de cr&te, la charge du choc et la duree de queue conllrme que les formes de courant sont similaires pes within the categories are similar, and this sug. dans chaque catagorie, ce qui incite la construire une gests the construction of typical shapes for each of these categories, forme typique pour chacune d'elles. . 4.2. - Construction d'une forme de courant. 4.2. - Constructing the shape. . Pour trouver une ordonnie i sde la forme moyenne To construct an ordinale T, of the mean lightning d'un courant d' impulsion de foudre, toutes les ordon. stroke shape all ordinates i, al point k on the se. nies in du point K des courbes enregistrhes. sepa. parate curves huwe been added and the result divided rsment ont st6 ajoutses et le rssultat diviss par m, by m. the number of ordinales, nombre des ordonnses mesurdes. On a do'nc ;
//ence a
I Yi in = mpl (2) t on m est le nombre total de courbes utilis&cs pour La d&tvrminution de T where m stands for the total number of curves con. u un point k. Ce nombre est tributing to T, in the point k, The number varies as varkshte du (nit' de la diW&rence des longueurs d'en. , registrement, due aux techniques d*enregistrement a resnts of different record lengths, caused by the . et de chiarage [3]. recording and digill:ing techniques [3]. , l 90004104 '
N 41 ELucina 35 TAH11AU IV - TAHIE IV Coefficients de entrilation entre les param?tres des impulsions suivantes nicatives. Correlation coefficients between parameters of negative following strokes. Courant de Durde de Vitene de .
' Charge de Dur6c de cr6te front montfe max. "Y Enerse l'im hulsion Timpulsion Peak Front Alax. rate o impu s charge "W ' "
current duration rise charge du ration Courant de cr6te 1,00 Peak current Durie de front 0,28 1,00 iront duration Vitesse de montic ina x. 0,11 - 0,49 1,00 blax. rale of rise Charge de choc 0,69 0,13 0,31 1,00 Impulse charge E 0,69 0,22 0,15 0,70 1,00 f.!"*'N" nergy Charge de l'im;4i " 0.43 0,26 0,28 0,62 0,54 1.00 Stroke charge . Dur6e de l' impulsion 0,25 0,30 0,44 0,37 0,12 1,00 Stroke duration - 0,05 l les coefficients en caracteres gras depameta la valeur entique au nse w de senificataan de 5 % - Coeffinents en br>ldface exeeed she ens,cas entue et ahe 5 % icvet of ownAconce. [)cgres de liberte : 73 Degrees offreedom 13 i Avant de faire la somme puis la moyenne des Before summing and averaging the pointss i it is ordonndes i,, il est n6cessaire de s' assurer que les vecessary to asce.rtain that the entves are proper /g i courbes sont convenablement ulign&es, c'est in-dire s ligned, i.e. the points i, in equation I2) should cor- l que les points i gde l'6quation (2) correspondent lespond to the same stage In the physical develop-au meme stade du developpement physique de la nent of the lightning flash. The sharply rising front l d& charge de foudre, l.c front raide des formes de .f the stroke shape is a suitable and easi!g recogni- ' courant est une caract6ristique approprise et faci- sable /cature in all strokes and the records were lement reconnaissable dans toutes les impulsions et therefore aligned in such a may that the 50 % am-c'est pourqual les enregistrements ont &th alignes piilvde points on the fronts coincided. .t different en faisant coincider les points d' amplitude 50 % technique, which ' computes the cross-correlation de tous les fronts. Une technique dilTJrente, qui frinction belmeen two curves and shifts one curve calcule la fonction d'intercorr61ation entre les deux by an anmnnt equal to the lag at which the func. l courbes et d6 place l'une des courbes d'une quanti e lion shows a .ma.rimum, produced an almost iden- I Egale au retard au bout duquel la fonction a un fical incan shape, proving the merit of the first maximum,'a aboutt h une forme moyenne presque (much more simple) method. In addition, all curves identique, ce qui prouve la valeur de la premibre were converted to a unit peak amplitude before ave-m6thode (beaucoup plus simple que la seconde). De raging. plus, toutes les courbes ont Jte ramenies A une m0mc amplitude de crite prise comme units avant de faire les moyennes. (a) Premitres impulsions positives. (a) Positive first sl roles. Dien que les impulsions positives soient caracts. AIthough positive strokes are characteri:cd by [. risses par des charges plus ilev&es et des fronts greater charges and slower fronts than their nega- l moins raides que leurs correspondantes n0gatives, tive counterparts, they do not have enungh common elles n'ont pas entre elles suffhamment de caract6- fratnres to produce un acceptabIc mean entrent ristiques communes pour permettre d'obtenir une shape. This muy also be due partly to the small forme de courant moyenne neceptable. G*est peut stre number of positive strokes which were recorded in d6 aussi en partie au petit nombre de coups de Ihr period.1 selection of 3 of the most Igpical of toudre posilits enregi tres dans la p&riode consid&- 21 recorded curves is therefore shown in Figure 11. r6e. C'est pourquoi nous nous sommes content 6s de 90004105.
M MScTRA N* ll I.2 Fig, tt
'O QR %,N
': 4 %
08 y.J % x,, 06 --* Q.s-
\ @**
O4 - \
; \ -)-D*
. Figure 11 0.2 [ N Formes de courant typiques - Impulsions positives
/I Typical current shapes - posstive stroken 00 0 80 160 240 320 400 4o0 560 640 720 800 ps 0 16 32 48 64 80 ps-8 0.0
\ ric. 12
- 0 ,6, \.
i
) --
g -- P ~a s l t / figure 12 )
\ ./^ '
,, / .e Forme de courant moyenne Premiires impulsions nigatives
- 0. 6 ] -- ' ,,
"~ A - ichelle des ternps 11 ichelle des temps
-0* 8 '- infirieure sup rieure 9 \ je* , . Menn current shape Nega*tive first strokes A - lower time scale il - upper time scale 0 80 160 240 320 400 ps-A reproduire sur la ugure !! un choix de 4 [les courbes les plus typiques parmi les 21 courbes enregistr&cs. 4 (b) Forme des premieres impulsions negatives. (b) Negative first stroke shape, f.a figure 12 montre la moyenne des formes de Figure 12 shows the mean neuntive first stroke premieres impulsions de courant n&gatives, tracht current shapes on imo different time scales (.\ and avec deux &chelles de temps dilT&rentes (A et B). B). In the region between 120-160 psec. the number Dans la zone de 120 b 160 ps, le nombre de courbes of curves cont:rbuting to the mean curve changes utilis&cs pour la d& termination de la courbe passe from SS ishc et and lonu recordinus) to 10 (only de 88 (enrchistrements courts et longs) b 10 (enre- long recordings) which e.rplains the ripple in this gistrements longs seulement), ce qui explique l'on- nrea. .\ further inaccuracy, inherent in the original dulation qu'on constate dans cette zone. Cn aulre recording technique (3), occurs at arov*rd 200 psec.,
d&iaut de pr& cision. inh & rent 5 la technique d'enre- most likely also controbuling- to residual errors here, gistrement emplo)&c b l'origine '3). se produit dans le voisinage de 200 ps et il est trbs probable qu'il est ici aussi une source d'erreurs r&siduelles. (c) Forme des impulsions suivantes nigatives. (c) Negalive following stroke shape. l.a ilgure 13 montre la moyenne de 7G tortues Flyure 13 shows the mean of 7ti negative following d' impulsions suivantes n&gatives, trac &es ici aussi stroke shapes. again on Imo Ilme scales (.\ and B). avec dens &chelles des temps (A et TU. Cn trait .\ striking feature of thest curves is the rather abrupt particulibrement fra ppa nt de ces courbes est la chunue in slope of the lait n/ter about J sec. and modification assez brutale de pente de la queue au the subsequent slamly decaying tail. Errors due to hout d' environ .ips, suivie d'une d&croissance lente the smne recording technique referred to above can 0 8 16 24 32 40 s-8 l rv. 15
-02 '
l I A
-04 _,__. Figure 13
-06 j, b- ~ ~"
~ ~l Forme de u>urant moyenne impulsionn suivantes nig4tges ;
g A - ethelle des temps li - echelle des temps aupe. ~
-08 I j " '" "'*
/ Steen current shape Neontive following stroken {
~10 A - lo uer time, scale U - upper time scale 0 l 20 40 60 80 100 ps- A 90004106 .
ll .,p 41
. . El.ECTnA 37 de la queue. Den erreura ducs h la m6mc teelmique be held responsible for the ripple at appro.riumtriq
.l'rnregistrement ausquelles il a 616 fait allusion 70 pscr. .lfter ubt,nt .30 pser, froin the sturl. acentary phn haut, peuvent 6tre tenues responsables de l'un. deteritorales sinec the entrent umplitudes in nonst of
,lulation olnerv6e nu% envirems de 20ps. .\u delk the original rectords were very low <>n the seule und
,lr .">0 ps apr6s le d6 hut, le trae6 devient incertain, et,uld not be urenralely resolved.
car dans la plupart (les enregistrements originaus, les amplitudes des courants h I'Uchelle adopt 6c sont tr6s faibles et ne peuvent 6tre lues avec pr6 cision.
- 5. - Remerciements. 5. - Acknowledgements.
l.cs auteurs remercient la Vondation suisse pour The authors thank the Swiss 1.'oundation for their l' aide qu'elle leur a g6ndreusement apport6e pour uenerous s"PPort of the tvork on .1/onte San Solen. i l'ex6cution de leurs travuux au .\lont Sun Salvuture tore, without which these data could not have been I et sans laquelle ik n'auraient pus &ts en mesure de cornpiled. Thanks are also due to the Council for l recueillir ces donn6es. Ils expriment 6galement leurs Scientific and Industrial IIesearch of South .lfrica i remerciements un Conseil de la recherche scienti. for financing the work of H. Kr&ninger who assisted ; ilque et industrielle d'Afrique du Sud pour le finun. In the astalysis of the data over the period of one cement des travaux de 11. Kr6ninger, qui a participe year. pendant un an h l' analyse des donnses. Bibliographie - References [1] K. HERGER " Die 51cucinrichtung (Qr die Hlitzforschung auf dem $1onte San Sahalore" Bull SEV, n*5,1955, pp.193 232. (2) K. HEltG'ER "Remitate der Hiiumessungen der Jahre 1947 1954 auf dem Stonte San Salvatore". Bull SEV n*9.1955, l pp. 405 456. [3l K. HERGER et E. YOGELSANGER "hiessungen und Remitate der Blitzforschung der Jahre 1955 1963 auf dem Stante San Salvatore". Buu. SEV., Bd. 56 (1965), n*l, pp. 2 22. [4] K. HERGER et E. V0GELSeGER "Photographische Blitzuntersuchen der Jahre 1955 1965 auf dem $1onte San Salvatore"
, Huu. SEV., Hd 57 (1966). n* 13. pp. 599 620.
~
(5l K. HERGER " Novel observations on lightning discharges : results of rescarch on 51ount San Salvatore". Journal of the FranlJin institute, vol. 283, n*6. june 1967. (6l K. HERGER "Slethoden und tiesultate der Blitzforschung auf dem hlonte San Sahatore bei fugano in den Jahren 1963 1971*?. Bull. SEV., Bd. 63 (1972) n*24, pp.14031422. 17] K. HERGEll "Oszillographische 51esmngen des Feldverlaufs in der Nahe des Blitzeinschlages auf dem Stonte San Salvatore'?. Bull. SEV.. Bd. 64 (1973) n*3, pp.120136. [8) F. POPOI. ANSK Y " Distribution de frfquence des amplitudes des courants de foudre" Frequency distribution of amplitudes of lightning currents. Electra n*22 (mai 1972), pp.139147. l l l 90004107 . t
, , R t F. 12. l l
1 LA L o f n i_ i f1 i sulated plate was mounted to which volt- I lYlCCnanlSm Or DreaKOOWn Or LaDOratory agewampphed sme a their data u ere
. obtained with a OKX 10t%ec Imicro-l second) wave which may for all practical !
r Pg purpmes, he regarded as a rectangular ) [ wave. The crest of this applied wave was j E C. F. WAGNER A. R. HlLEMAN h * 'd '"" "' "' 3 4 5 '"' ""d '
- ffLLOW AiEE MEMBEit A!El the gap were varied bymg chang?"di'i""
the gap l length by raising or lowering the sphere. The current Dowing from the sphere HE AUTHORS' primary interest in Later a more detailed discussion, sup. was measured by means of a cathode-ray 1 Tthe mechanism of gap breakdown isinby experimental evidence, of var-ported oscillograph and the setup was housed
, its relation to lightning stroke phenomena. ivus phases of the dinharge will be under. in a black box to expedite photographing In considering the efTect of the lightning taken. It has been known for a long the discharge.
stroke upon transmission lines, the prob- time that the breakdown of long gaps Fig. I shows a replot of the arrent j lem has been resolved into two parts. proceeds in several phases. These are from the sphere as this substantiallv rac. The first of these is the electrical response clearly de6ned by Park and Cones. tangular wase was applied to gaps of
].
of the line to specific assumed stroke Their setup. Fig.1(r consisted of a various lengths. Some liberty was taken characteristics. A methodP8 to deter- metallic sphere of 1.0-cm (centimeters) by the authors to line up the large current ! mine this response was recently presented diameter mounted on a vertical shielded pips as this illustrates the phenomenon ! before the AIEE in which it was shown rod connected to ground which permitted more clearly. Small distortions of time l { that the time change in the charge in the measurement of the current from the scales result from this p ocedure. The j
; stroke above the tower may be as im- sphere alore. Above the sphere at a time of application of soltage is in- I portant as the current fed into the fhed distance above ground a dat in- dicated by the small curr.nt ripple just j transminion line. The second part of l
, the problem is the determination of those ,ps g a t A
spsEng l
.i stroke characteristics that are required to NEG A TivE Po SITivE implement this a pproach. An initial L.sccm e ffurt' als.ng this line was previnusly pre- __b . __ g _ ,
sented in which an attempt was made to g
,[,
synthesize the stroke characteristics by L=46cm correlating the known stroke character-g, , g f p m istics with laboratory determined char.
& L*di H -
acteristics of long sparks. A further j effort
- was made to utilize the available _4_
~
information concerning the electric l' eld ' ' 36 l C produced at remote points to determine u
, the wavefront of the stroke cu rre nt. L = 25 cm }
The purpose of this paper is to present a N 9-review of available information on labora- '*30 ! tory produced sparks, to present new data on this subject. and to co ordinate the external manifestations of gap break-down from an engineering point of view. o
*77 L 20 cm ! L = 25 These results have been co-ordinated in a companion paper.' in this issue, with
) j v CaoSECONDS similar data from natural lightning. 3 o 10 20 L ' --- --
General Phenomenon E
- t. 2 o 'f FIRST CURRENT P:P L = is em s /
In order to establish the general nature ' N of the phenomenon the experiments of f sEccND DISCmGE j i msE Park and Cones' will be described first. a I Paer 614tio. recommended by the AIEE Trans- l M minion and D:stribution Commutee and approved by the 4 t h E Technical Operatiurs Depart ment for presentation at the AtEE South E.s st - Sout h F
] { k Central rastrict S!ce tin g. New Or cans. La., (
Aptd 5-7,19%. Manuscript submit t ed January 9,19G1; made available for printing March 3.1961 L
- 13 c m { .1_ Q 86cm I6cm- ,
C. F. Taox en and A. R. Ilittua w are both with ' the Westin g hou se Dectric Cur pir at ica , F ast o
~
Ms gh. Pa. 10 20 The authors wish to ack noaledge the valu able YiCAoSECCNOS 3 en. stance of P. IL Long, J. Prado, and L. K4'ning c e 3 tem ( n for performing the gap breddun tres and of
"' fi S
. YeEENItio$'fUpTd ann 7dr'as'e',"i!r""nN 1- Cu"*n!8' 5'!' int *h'a *a 0 07 X100+$'c 16k' I'nPO' i' 'PP ii 'd th' DP
} nahon of the gap time lag curses. $ttup el 5bo*n in M G04 Wagnct, Hilcman Ifa:icnism of BncMown of Luomhry Gaps OcTakn 1DG1
record the final current as a protective rise starts. This change consists of one preteding the 'almost vertical current or more bright dischar;e channels that change which represents the chargirg gap short. circuited the oscillograph ele-me n t. This rise in current has been start at or near the sphete. These dis-cunent to sup;ly the electrostatic f eld chargas have been termed " channels" by before the gae us di,ctr,e phenernenon termed by Park and Cones the "second dischar;c ri.-e." Park and Cones to distingui-h them from occurs. The large abru pt char;c is the fir =t kind of diwhar;e w hi h thq termed the "first current pip" by Park The tirst current pip is appr aimately and Cones. Generally the current pip the same for the application of a ne;2tive termed " streamers " There channels impuhe wave, except as to ma;nitude, as mme in zipaq fashion across the gap, does not coincide with the instant of The length of the channel as a function application of voltage but is somewhat that obtained for an applied positive impulse, but the second dischar;e rise of time was measured from the progress dely ed. The delay is ocea ioned by the of the photographed tip and is shown in is not as gradual in its change and has chance occurrence of a free electron com- Fig. 2. ing within the overstresed regivn in the steeper rates of rise. In each case after the second discharge rise has begun, com- To form a working thesis on which to vicinity of the sphere and triggering the plete breakdown follows unless the volt- base discussion the following explanation dN har ge. This is usually called the "stati stical time delay" as it depends upon age wave is chopped. similar to Park and Cones' is otTered to describe the phenomenon. There is no a chance occurrence. The crest of the PostilvE DISC 11aRCES dr-t current pip for the sphere posttive space charge in the interelectrode space deceases from 1% amperes for Le-22 cm The photot;raphs exhibit different char- prior to initiation of the dischar;;e. The acteristics for positive and negative dis- charge on the sphere is determir.ed by the I as the gap is increa-ed and for the sphere l charges. Since the discharge is the sim- electrostatic solution for the particular negative from 20 amperes for L = 10 cm. The magnitudes vary over wide limits. pler with the sphere positive it will be configuration. The field is strong near The current rises to crest ;n about 0.01 described first. For the longer gaps the the sphere and decreases as the plane is photngraphic evidence indicates that approached. If the gap spacing is suffi-uw and then drops somewhat expcn-evialle to zero in about 0 3 gee for the streamers that produce light radiate from eiently small, a zote in the vicinity of the the sphere but do not cotaplete the pam spi,cre is strested beyond the critical value s;.here either positive. or ne;stive. The current ceases af ter the disappear- sage of the. gap. As the gap is reduced of about 30,000 velts per cm. As a free ance of the first current pip at larger gap to 2 3 cm some of the streamers do complete electron appears in this overstressed passage but the current still drops to zero zone, streamers form which radiate from settings bu't with the smaller gaps, such as L = 20 cm. for the application of the posi- and breakdotm does not occur. Ft. L= the sphere. The streamers are not uni-tise potential Fig.1(L) shows that after 20 cm complete breakdown occurs. form in length and with a suMciently large the current almost decreases to zero it Photographs taken when the wa,es were gap none reaches the plate. The cet slowly rises again becoming more rapid clupped by means of a parallelgap show etTect of the streamcrs is to produce a and is finally limited by the constants that the appearance of the discharge be- space charge that develops its own tield of the circuit. The oscillogram does not gins to change after the second discharge and potential drop. Since the applied volta;e across the gap is constant, as the space chan;e develops less potential is j 3 _ available to produce the charge on the j { l i sphere. Consequently the eb-tric f eld j {$ 3 ; a, .___L l 1 i I adjacent to the sphere dectern until it reaches 30.000 volts per em at wuch poir,t I % i ' further supply of cunent is inhibited. j ! l The net result is that when the gap is
*> is l
l l nl-l long, a ball of space charge forms around l ; the sphere. A state of equilibrium is at-
! l t l -
tained for which the formation of the i* , space charge is arrested and ionization l fC
- o processes cease. If the gap is decreased 5
* '
- and an impulse is reapplied a new state of I" j I equilibrium is attained for which the ball r j !
of space charge is larger. Finally, as the y, i I o J_ gap is further decreased, the space charge
+ I 5 l l j l expands to occupy the entire interelec.
trode space and, with additional reduction, 7m i , __ [L._ l e nditions become conducive to the de-5 t.t N ou velopment of a channel from the sphere.
'N l l1 4- The discharge is converted from a gluw ou~ s -- ;
o 1 I discharge to an are plasma that begins l i
! i /! Fig. 2. Progress of to grow from the sphere toward the plane at a rate indicated by Fig. 2. The
+ * ~ " channel tip across a 20-em sap' fo' criterion for the critical condition, next N'
a a e ~ L. - l
] - ]> f v t toci t y the conditions of Fis.1 for the sphere to the sphere, appears to be such that a charge distnbution wd. l produce an aver-8
~ ~ "' positive The dotted l
~
line is the slope of age gradient across the entire gap of about ' f < 0 - the distance-time 6,000 volts per em. Conditions govern-0 $ o, 3, ,, m ,, 3 3
'we m wiu ostcoa.cs (solid line) c urv e ing the transition from corona to an are 60 *> ;
OcTouna196L 11*agur, Huonan-Meclianhm of Breakktn of Labomtory Gaps 90004109
i i k E g , . p..-., ~2eo
'& m w *
.M.
av 2co: 3
.A c, J ~i
-ti f g w .[ -
B 5[ m*
- 2. -
- w. 9...;.g;n.. p, D A M.
w% c t. -,so u. jer, t:- *
's o c- .,-, e vt%. s ; M,,,gg, . . .s W. am w '
/
- p. : -ic y sm 'I l~ry..$* \ >
@R My Q:&_ f. MmJ .h' q ,a.f a gnm *. MW r*rg. Q [ W -. .",
i a.u.a - f&, / ,j . so SOC - .h . ,p. ' a a' . te .. a e* v n : 7 .,g . ]8 ? j 3 y& Y L a f M " t' I" t c : A =3 i_ .o (a?%CfMW<r
%'. w b = c .. wnw M. Qk. >h >c m;*q;
\ n
* * \
~
s a. < n
$f #N C th i'* o. S lo is 2o 30 u nosev os o s 10 Is 20 30 m a n e w.os p$&jyj.MM"LNh-N% * *f0:yW y9.Q: : 5,7."'":. q cc
- ~ - c gra -,so '
u c-en}; W .. n.f .:.,a A.w;q%., . % ' mo n C ^: ev u: O . N.c5mg 6.Q M4*hm si / ' , k * - j .- . F
-, oo, f- '{} I . ( - -
-h- - -
' 00-M j I{ n
- 3" l .M , . c-[t i
i, mo$ 'm-Y,( G_x,;LfA' - r S.,_ , _ _._m., co-
- 3. . . dI ~ o. - .. -o
. e
$, < . : w.d F
.~ , -v. c. < n
- qp. 5-'_- , . . . .p .: ,, _a, t*
' %' ..'.'.' .T. T* .7.. J. 1.m
[ . W%fh 0.UQQ$%
~"
* ~ ' "
.l h! SIS %Y W%'.h$'d.d,W .M'NI2 - !' .o- $: $ A E la- 'o s io mcaostconos-
[s 20 30
- ,. q cnostcows N
l @ T&yf.&fdys J1 - - -
" NC,~,r $ <[MNdh >
.. a fig. 4. Sparliever of a 16-unit insulator string equipped with ccing rings of 4-inch. diameter pipeu b . . . y .- . - i s ..
f%. A-Five successive epolicadons at increasing voltage h' a% _ T g;r. O,y,'j-Y@M j _(.1p-Qf
,' i { g[( B-Currents resuhing from A C-Soll higher vokase with corresponding current
[ D, D-Apphcotion of vohage and resulting current with the Msulators onk
$ ....-.'f- Er:. p' ', OW+~Q7 ' ,. Nele the sudden nse of the current prior to :he !aur increase L.O. :.J.%1.k, ..N.:y. l?hw da.JA c-p% Qr w ry,s e b%N~m&
ft c - -69xc .wc n*8,7.',k/ - mechanism of production may be quite in mid-gap. From this point the current different. The negative discharge is is again determined by the constants of the C 3. Current responsen of a 900-inch more diffuse than the pusitive i surge generator and the characteristics of rod-rod gap when a 3 X50 usec 3,000,000- . volt surge is applied visually and photograph. wally. I t u .,. the arc. sists of' numerous streamers of fine texture OAAAA14 A A--Vehase wave while the positive discharge consists of Disct ssm" 8, C, and D-Current waves 7 U U ll't I l U
, fewer but stronger and more crooked The general nature of the breakdown
. E--Timing wave of 100,000 cycles per second streamers. of nonhomogeneous gaps, consisting of a j in the authors' interpretation of the corona discharge followed by the de-i are not well understood. It is neces- phenomenon just before sparkover it is velopment of a conducting channel, has sary to wait for further developments by assumed that the electric gradient, be- been knea n from the earliest davs of im-the physicist before a more definite cause of the space charge, is approxi. pulse testing. For example, Slepian and explanation can be gis en. The drop mately constant across the gap although Torok'in 1929 by chopping impulse u aves along the plasma channelis sery low and this assumption is not essential to a showed by means of photographs the thus when it is initiated the effect is generalunderstandingof thephenomenon. stages of progress of the discharge and progressive as the channel merely con- For the authors' purposes it is evnvenient also some indication of the maximum stitutes in effect an elongation of the posi- to think in terms of an average gradient curre n ts. 1*tilizing a rotating camera tive electrode. The continually decreas- across. the gap. Returning to the Park Allibone.' in 1938, presented an extensive ing gap length encourages all the factors and Ccnes exp 'ments, as the gap is study of di< charges in long gaps and es-originally responsible for the development decreased and . 3utriciently small, the tablished the chronohigical sequence of of the discharge. When the head finally gradient in the interdectrode space the leader followed by a return stroke in reaches the plate, the channel constitutes due to the space charge increases and the laboratory. He commented upon the a virtual short circuit of the surge genera- the charge density next to the sphere absence of a discharge from the plane tor and the subsequent current is de- is higher than at the plane. But the when the rod of a rod plane gap is posi-pendent upon the constants of the gen- conditions necessary to initiate a charmel tive. This distinction betw et u the corona erator circuit and the characteristics of from the positive plane are reached prior starting < oltage and the breakdown was the arc. This effect is illustra'ed by the to the conditions necessary to initiate a recognized as early as 1931 by Cmdiet, current oscillogram for the 20-cm gap of channel from the negative sphere. So Edwards, and Perry."
Fig. L the channel starts from the plar.e before Komelkov"in 19& working with gaps one starts from the sphere. But as the between lo and 100cm concluded that the NrcArive Discuracts pc sitive channel progrenes from the plate, drop in the ihannels u as ury low, about With the <phere negatise, the charge conditions beer me more critical at the 74 vats per rm, and that the p&nt develops within the interelectrode space sphere and a channel is fmally initiated ir the corona streamers was in the urge in a similar manner, although the actual from there also. The tuo channels meet of 0.000 to 10,000 volts per cm. Saxe and l CO6 ll'agmr, Hileman-%chanism of Breakduwn of Laboratory Gaps Octm.n 1%1 l i
O .\ leek'" als.) concluded that the drop along some manner produce an electric field that available to produce the space charge. the cha nnels is small. inhibits further growth of the discharge. For these cases, when the gap is ad-IIag-nguth, Rohlfs, and Degnan'* justed so that the corona space charge furnished limited evidence on a vaster AvEPAGE ELEcTRtc GRADIENT AT envelops the entire interelectrode Space, geomectic scale which might be viewed as SPARKoVER the 3pplied voltage divided by the g3p supporting the geretal nature of the For Positite Discharges length is the average gradient aloog the discharge discusped here. They measured axis i the corona envelope. As has been mentioned, Allibone com-the cunent dowing in the ground electrode rnented on the absence of an upward Once the channel has begun to form the of a 200-mch rod-rod ;a p w hena 3.000.000~ channel from the plane of a rod plane discharge develops to ultimate sparkover volt negative 3X'io w see impul e was ap- gap when the rod is positive. Park and almost invariably. For the critical spark-plied to the free electrode. Fig. 3 ts a Cones also obsened that for the sphere over voltage, half of the applications of reproduction of Fig. 12 of their paper. positive the channel proceeded from the voltage produce sparkover. Conse-Curve .1 shows the applied voltage which quently, this voltage constitutes a meas-sphere completely across the gap. Prob. for this gap was just below entical. % ith ably the absence of a channel arising from ure of the average gradient of the space
.10 applications of tius vcitage, nine cases the plate was most dramatically con. charge to produce sparkover because the developed a glow that bridged only a firmed bv Norinder and Salkau in their cases that do not cause sparkover repre-portion of the gap and the current m the '
elaborate photographic investigation of sent the maximum development of the grounded electrode was as shown by spark discharges. The absente of a space charge without the formation of a trace Si in nine cases the glow bridged . . . C >ad" Gel. comph.eation caused by the formation or a the entire gap and the current was
. channel from the plate insures a some.
In Fig. a the average critical sparkover as shown by trace Ci and in two cases what simpler analys. is for this type of dis-gradients of rod to. plane and rod-rod complete sparkover occurred and the charge and for this reason the rod-plane gaps from different sources for both current was as shown by trace D. The discharge with the rod or small sphere as polan. .ttes are plotted.
. he posittve magnitudes of the three current pips .
positive e!cetrode wd. l be considered first. pelarity data are indicated bv the full were remarkably consistent and avera.;ed h.nes. The Bella:.ch.i and Tea.;ued data After a vpttage is applied across a gap. 9 amperes. The duraton of the pips the space charge expands and becornes represented the full wave (1.5X40-usec) was 9 usec. m teintense until the gradient next to the entical sparkmer values and were made As early as 1929 Torok and Fielder" l electrode drops to ",0,000 volts per em as on gaps up to 200 cm. Breakdown oc- i measured the predischarge currents of was ment.ioned eath.e r. When the rad.ius curred at about S usec. The Hagenguth suspe n3 ion insulators. Fig 4 is a re- of the rod of a rod plane gap is small, then Rohlfs, and Degnan" data covered an prod um.on of some ;f their oscillograms. even greater range up to 640 cm with an n y a sm (.mt Umit mo, for a In all nf these records the negative pole Pointed electrode) potential is required impulse wave of 3X50 usec. They stated was grounded. Fig. 4(D) shows the t Produce a charge on the electrode further that their unpubli3hed data, with voltage and the cunent for flashover of a result m a gradient at the elec-gap spacings up to >0 feet tended to give strinz of 16 insulators. The delav in cur-rent occasioned bv the necessity of the r e of s per ent en pr p amage gradients of the same value. j
~
positionin$; of a free electron tically all of the applied voltage is The Gorev, Zale>ky, and Riabov" data, correct is evident and the resultant current is typical of others that have been presented. Figs. 4(A) and 1(B) are mates; the i gl former shows the applied voltages arid $ "o -
\ _ g c o sa, ,% ,,, u s the latter depicts the resulting cu: ts e 3t% ,, m aw ec3 w s l 3
(( , , as a 10 unit insulator string with an arug = / @ pruam au mad' , rmg of Linch pipe was t' ashed over. Fig. E y o. V,./' , - p yt, ,y 3,,, s i 4(C) shows the same test piece with the E '. ' ,' '1, C3 application of a still higher voltage. The short circuit current of the surge genera-3 z
'\ O l ,^\;
\
en ro n.are
; \,.( ' .. , ,,,,.,x 3 ""'
tor was mn4t likely about 2,000 amperes. \g ' i 5 @ 5mE5
- 22** '3 5e$**
It is not known whether the current trace recorded the true maximum as it $ K - MEO fo WE wD a3E GRAD EUs M r *E sa %8 E <aal" may have been limited by the operati..n z j w_ %yf
' t f... @ 1 prL.," 4 [
of a protective applaced act etheshunt l
'.,.../ _
Thus with this general background of ; , ,__ , the phenomenon ar d hnuted historical revie w, a discussion of the component 2 f * *' c__. s T ./,,,"' - ____ y,% , ,c _,, , ; l parts of the discharge will he undertaken. y ~C -
~
o w., f' x
\I s
\t
~ y z'
a dj Corona Strearners and Envelope It is clear from the foregoing that for No & 77'r'7 90004111 ct 4 n1; impulse voltages in excess of the curona .ooo - . , threshold voltage butless than the critical '" 23 2" GAP LENGTH lN CM
*" 5" 5"
]l breakdown value, a self. limiting space char;e is distributed thruughout the inter- Fig. 5. Averese crest impuhe spadever gradient for bcth positive and nesative polarW Ior electruie space- This charge must in saps of dUferent conas uration as a function of sep lensch Oci m a1961 Wapu r, Hileman-Mechanism of Breakde:cn of Laboratory Gaps 607
s 6 , with 1.5X40wsee impulse waves, agreed the anode. During the initial stages of 1 indicates, the time to reach zero is about quite cleely with the other data. The the channel growth the velocity of the 0.3 usec. The naveshape stays essen. Norinder 'and Salkau curve for tFe rod- channel is small as compared with the tially constant for both polarities, for all
]J plane sparkos er was only 11To below that later stages. As the head of the channel gap spacings, and for fast and slow ap.
l of the others. The curves marked by progresses, the space charge tends to plied voltage waves. a small horizontalline with curled-up end devCop ahead of it. The initial channel Remembering that in Park and Cones' f" t represent horizontal wire.to. plane data. They agree almost exactly with the rod-development from the anode merely experiments the applied voltage was kept ser es to shorten the gap length and in- constant and the gap et uditions were
! plane curves. These curves indicate a creases ht e gradient adjacent to the cath. varied by chancing the gap length, the practical wo-king average of the average ode to a point where a cathode channel fully developed discharge will be taken as crest sparkover gradient for long gaps of will form. The influence of the anode that for which a 0.0TX100-usec positive e about 5,400 volts per cm for rod-plane channel on the cathode channel should wave produced 50To sparkovers. From gaps and about 6,000 volts per em for rod- be proportionately less for the long gaps their data this was 24 cm for positive rod gaps which can also be interpreted as than for the 111 gaps and shculd be polarity. The effectis e vek. city of charge the average gradient alung the axis of the less for a plane anode than for arodanode formation for the peitive space charge, gap produced by the space charge. So, in order to estimate the avera;;e r,* will then Le dermed as the velocity ob-t Part of the discrepancy between the g,cadient at which channels are des cloped, tained by dhiding the half gap lergth by J laboratory data may be ascribed to differ- one should refer to the sparkover data the time required to produce the fully l encesin waveshapc of the applied voltages for rodplate gaps for long spacings, and developed field. Thus j as well as differences in methods of im. make some allowance for the develop- ,y f pulse measurements and to laboratory ment of the channel from the plate. r, + = #* "
h and observational conditions. Berger's From Fig. 5 the critical average gradient x0'3x15 h CIGRE reportJ8 which discussed com- at which the negative channel develops = 0 0013c (1) 1 parative tests by 14 different laboratories, is estimated to be 8.000 to 0,000 volts /cm. For the negative polarity, the effective
, mdicated that the critical sparkover vdocity of charge formation is volta;;e gradient for a 45-cm rod-rod gap TIME ;o EsTABt isH rHE CH ARGE r
for positive polarity gave a range of from 11.5 t pWgraphs of Ge paddags b. ~ 6,000 to 7,600 volts per em even af ter U x 0 3 x 10-* air density corrections had been taken into (used merely to apply to that which oc-curs before the production of the conduct-4 00W W conwera tion, ing channels) show very pronounced The paper by Hagenguth, Ruhlfs, and streamers of high light intensity. The Degnan" prm ided another factor. At For A,eptire Discharges head of some of these streamers travel at the critical sparkover point of a 000. inch In Fig. I the waveform of the first very high velocities. For example, Park rod gap with a negative impulse applicd discharge pip has appmximately the and Cones stated that the mean streamer to the free rod, the time required to
, same waveshape for both polarities but velocity was found to be 500 cm per develop the space charge was about 9
; the magnitude for positive polarity is usec or 1.77o the velocity of light for the usec. This corresponds to a velocity of j somewhat larger than for negative polar. sphere negative and S00 cm per usec or 200x m j ity. A correspending difference in the 2.3To the velocity of light for the sphere v.=, p = 2 SX 10' em per see J relative charge fvr positive and negative positive. The average deviation was 90 applied volta,;es supplied to the corona for the sphere negative and 100 for the =0m W envdope of a cylindrical conductor above sphere positive. However, these nu mbers Considering the wide range in gaps, from f a plane when impulse has also been noted cannot be viewed as the actual rate at 4.3 to 000 inches, to which these valoes by u*agner and Lloyd." which charge was developed in the inter- applied it is remarkable that these num.
As mentioned previously a funda- electrode space. S;me other mechanism bers are so s ery nearly equal. mental difference does exist when a rod- must have been present which the physi- Concerning the actual physical process plane gap is impulsed by a negative po- cist may help to explain. As mentioned invoh ed in the establishment of the space g tential and when impulsed by a positive by Park and Cones, the streamers "should charge. it will be observed that the elec-
; potential. This difference is the appear. be thought of as a traveling wave of high tron drif t velocity in a field of 30.000 ance of local discharges at the plate after charge density which is propagated by a volts per em and a pressure of 760 mm the negatise corona space charge has process in which new charges are con- (millimeters) is, from Lneb,5 about 1.4 developed to some extent. One of the tinually produced at the leading surface X 107 cm per sec cr 0.0005 c. This field plate discharges finally develops into a of the ball by the high gradient there. is chosen for comparison purposes be-
, plasma channel that grows toward the In the path behind the ball there is left cause it lies midway between the initial
{ rod before a channel develops from the a high concentration of both positive and final fields. This value compares rod. Because of the presence of the and negative ions, with an excess of favorably with the values given by equa-l discharges from the plate, the sparkover pmitive ions in case the sphere is positive tions 1, 2, and 3. j curves cannot be used directiv to deter. and an excess of negative ions in case n 'A 1 j 9
, mine the average gradient th'a t leads to the sphere is negative." The shape of CHARGE AND ELECTRIC r1 9 V t i I L.
i development of a plasma channel from the the current wave of the first discharge DrsrRIst T1oN % rTHIN TnE CAP j rod. One wonders at what value of pip is quite repeatable for the sphere So far consideration has been given to j average gradient would a negative space positive and somewhat less repeatable for characteristics of the space charge that are charge develop into a channel from the the sphere negative. The average wave- subject to actualexperimental determina-cathode if the space charge were permitted shape rises to crest in about OSOS usec tion such as the averagecriticalbreakdown to form from the cathode without inter- and decays approtimately exponentially gradient and the external current feeding ference of a corresponding discharge from to half value in alout 0.0S usec. As Fig. it. Because of the dif!icultiesof measure. GOS Wagner, Hikman-Mcchanism of Breakdoun ofIsboratory Gaps OcTo n n 1DGI N
L f, ment, little is known of the actual Fig. 6. Wlocitr of the teadir cy g structure of the charge distribution or hlopment u a function of . - the terminal voltase, ui, for c x io' of the 6cid distribution. Doubtless these n distributions are a function of time. thru dhent constad volun ocom - of the unbridged sep for a "i
- I Photographic evidence points to the early j
development of streamers which may be quite independent of and unaffected by
, {#, , , *[ .
em end series rutstance in the * # each other. They probably are responsi- circuit is 2,000 ohms ,
,,ococ, ble for tomzation phenomena that pro- "
duce charge separations. The speed with ,
* {
which the space charge develops suggests 5 . strongly that its development is asso- $ * * ^ ciated with the movement of electrons *' ,p ,, c, rather tLn ions for both polarities. In . time the cwvements of these charges pro- a j 0
- duce a mass or aggregate effect in which ;
3 all the streamers play a part. In contemplating the average electric gradient just prior to breakdown, from h j go j .aoj fo v,60m ago v ma Fig. 5 for positive polarity, one is imme- , ,3 ,8 ,' diately struck by the fact that it is con-stant over a very long rangeof gaplengths. What sort of charge distribution would , with a 200incn gap. With slower rates then more rapidiv. According to Park give rise to an avera.;e gradient that is of r sc and an abur. dance of electrons and Cones, at midgap (10 cm) the rate of , mdependent of gap length? Park and to trigger the gap, the current supply- growth .is about 20X108 cm per see or Cones suggested a charge concentrat. ion ing the space charge is reduced in magni- 0.0007 c. that varies inversely as the distance from tude and spread out over a longer time. Akopian, Larionov, and Torosian" the sphencal electrode m their sphere- A considerable gradient exists within undertook elaborately combined oscillo- f plate gap. Fyr either a truly sphencal or & velope and for the posi- graphic and rotating drum photographic truly cylindneal charge distribution the . tive discharge. Just prior to sparkover, the tests on rod plate and rod. rod gaps of resultant e!cetne gradicnt is a constant. va is about 5,400 to 6,000 volts per 100-200 cm with positive impub.e poten-For a rod-plane gap with the rod positive' em and about 8,000 to 9,000 volts per tials applied to the gap. Thus, they were such distr:bution can be siewed as bem_ g able to co-ordinate the travel of the head cm for the negative discharge. produced by the following mechanism. * # """# Supp*e that posithe ions and electrons anneb vane M oc ternund dag home
**' "Y'"b are produced uniformly along the nu- kov" had previously demonstrated that merous very.high-speed streamers that It has been obsened by a number of the drop in the channel was about 50 volts l emanate from the pos.tise i. rod. As the ..
mvestigators that the positive discharge per em. Therefore, assuming the drop electrons move toward the anode, if the
** ane gap a mu w be negMMy smag the dage acmu ,
posithe ions that are left behind have a more stable and em.s. is tent than the nega- the unbridged portion of the gap is I-uniform radial distributen along each l tive discharge. This apph.es particularly .dentical i with the terminal voltage. streamer, then the volume d.tstnbut. ion to the development of the channel. They showed as indicated in Fig. 6 Wig. r would vary mversely with the radms. Probab!v this explains why more data S of Akopian, et al.) that for a rod. plane , The magnitude of the space charge y,g . g ;g g current and the photographs of the dis- nne , ve y heaM oe channel a ' I I charge both indicate that the mechanism for a constant value of the unbridged g sp Posmvs CnmEI.S of the negative discharge differs from that - vaned Unea% wig 6e apphed dage. of the positise discharge, but the result- I, ark and Cones 7 presented the data In this Sgure u, is the applied voltage m, 1, i ant char;;e distribut. ion may still result t concenng Ge su M .is in a 6eh! that is substantially constant. s own m b,r.a ge gap tergth in cm, and x the head of the bn ghtly luminous posi- the unbneged portion of the gap in cm. i tive channel as it moved across the gap Cun es are drawn for three constant va!ues 1; ImRnt St;sotany of the unbridged gap The velocity U of the setup mentioned previously. The ' It appears that if a rectangular voltage gap was set for 20 cm and a 0.07 X t, wave rises from zero at a value of termir,al is applied across a nonuniform gap whose having a crest magnitude of 145,000 volts voltage u, that would prodm e discharge j
. average gradient is just less than the was applied which was choppcd by a when applied for some length of time , J values given in Fic 5. a self-limiting space parallel gap. The symbol I, indicates the (" prolonged action" according to the "j g
charge develops that inhibits further flow time after the nrst current pip at which language in reference 21). Beyond this of current. The flow of the char;;e the wase was chopped. Corresponding voltage the velocity is proportional to the excess of the terrninal voltage abose this I ]a into the gap is at a rate of about 0.001 c phott aphs of the discharge showed which corresponds approximately to the the distance that the channel had pro- value. The values of u, for the three electron drift. Park and Cones' data gressed during the chopping time. The cases are indicated below the ab3 cissa. t show that for 10- to 20-cm gaps this slope of this curve is plotted by the They also showed that the positions of the , , development requires about 0.3 usec for dotted line and indicates that the straight lines are related and that for L 3 its completion and in the IIagenguth, the initial velocity is 3X108 cm per sec rod plare gaps the following relation for > Itohlfs, and Degnan data about 0 pec or 0 0001 c which rises slowly at 6rst and the velocity holds. 90004113 609 Octouen 1DG1 Wagner, Ildeman- .llaltanisrrs of Breakdern of I.aboratory Gaps
=k es-se cm per gsee (4) from the rod or <phere and met approxi- that the velocity of approach of the z-0 23x mately .in the truddle of the gap. channel ups is a function of :he instanta- , For a rod plane gap up to 200 cm k is Roo-Roo C Ars -
about 9 and for individual discharges the f the unbridged gap. As a basis for his coetlicient k may diverge from its mean For rod-r9d gaps, channels form from work it was necessary to determine ex-E value within *20To. With u, known as la th eledmdes. Abpian, Lanonov, and perimentally the time to sparkoser of I a function of x'(very nearly linear), the Tomsian determined that for an electrode gaps to a rectangular applied volta;;e [' velocity v and consequently x can be separati n of 125 cm the velocity with wave. He found that by applving such sch ed in terms of the applied voltage n,. which the channel tips approach each a ware to irradiated gaps from 10 to 70cm i Akopian, Larionov, and Torosian" hase ther can be expressed in the relation the t,une to sparkover, r, could be ex. tested this procedure with applied voltage u , - u. F'"' F ' "*E """ waves of widely difTering shapes with """ C
- W " W '
x gratifying results. where u,is again the actualirstantaneous
' " [\2910@.U / * * *C ho A nvr CmNEt.S volta;;c in kv across the electrodes and where U is the mapitude of the applied
. The negative channel is much more u. is me critical vo t ge in ky rf the rectangular wave in kv and s is the gap
[ rrratic than the positive channel but unbridged gap, x in em; They general- length in cm Rusck also stated that be-i because the experimental results are ized no further than this smgle gap but cause the enocal sparkoser voltage, b, I usually complicated by the presence of did show that tius relation produced is approximately a linear function of the
# g od results when the applied volta;;e was distance s, the time lag can be expressed positive channels, it is difficult to dis-criminate between the effects of the two varied over a wide range of wavesbapes, by Rusck," on the other hand. stated that polarities when both are present and in a g
<!e velopmen tal state. Examination of qu.s appreach was not a complete solu- r=U p in rec (7) i t he channel currents of Fig. I reveals that wn because kus mde in Ns labratmy how that the formula given ,m the above His relatit ns were satisfactorily utilized
; the currents rise more sharply when the
, spliere is negative. This may posibly mentioned paper caimot be utilized on for ditTerent types of applied waveforms.
. indicate a higher vducity for the thannel er g ps." He deverly obviated the He w arned that his work should be applied developmg from the sphere. It has also amp of taking photographs of the to time lags less than 4 to 5 eec, as in.
heen obser.ed photographically that the barge by simply accepting two im- correct results would be obtaired for positive plate t hanrel progressed a con. ponant awnpuons that are also m- longer times. siderable pertion of the gap before the herent in the work d & pian, Lasnov, negative channel started from the sphere. nd Torogian. First, that tbc drop in the Observations by the Authors Dut in spite of this handicap the two channel is negligibly small and conse
- channels met in mid-gap. This was pos- quently the channels can be viewed as If u,in equation 5 is explicitly ddined sible only if the negative channel traveled extensions of the electrodes, and second, as the critical sparkover voltage of a with a higher velocity.
Similar evidence has been provided 2s by the experiments of Hagenguth, Rohlfs, and Degnan" which were described pre-viously in connection with Fig. 3. In ~ nine out of 20 shots with the same l voltage applied, the glow bridged the g . _ entire gap nithout sparkover. In Fig.11 o ;m ges a,
<>f their paper a well defmed streamer can he seen " progressing from the grounded s' 3" C
Q ". '. p o mis - positive rod within the glow emanating from the negative electrode." For the , particular photograph shown this streamer
~ ~
2 has progressed about one fourth or one- 5 , , *csmvt third the distante across the gap. "On P - complete breakdown of the gap (not < ** d shown) at the same voltage there is a {g = i well. defined split in the spark near the j io - { middle of the gap, indicating where the the final streamers {in present terminol-
, ogy, channels] emanating from both elec-j trodes, met." This experiment also
, strongly indicates higher velocity of the c3 _
j negative channels. Norinder and Salkau related rimilar i experience with rod and sphere-phte gaps. The plasma c hannels began at the %1 Tirne-W ms b plate (anode) and proceeded toward the sundard md md ws in re-sponse to a 1.5 X 40 usec volt-90004114 rod or sphere electrode. At a consider- l ese
- eve b spacings from o 2 . 6 e ~f T i ~ Ts ably later time sintilar channels emanated 20 to 100 Inchesu wt w wcaest cms p tilo il'agur, Hi!cman-Mahanism of DreaMan of Laboratory Gaps navnu1%1
@ -- - - - it forms a gw;d conductor extending as a thin pened from each electrode. In order to satisfy the condition that the 35 [
electric gradient alnn; these good conduct. ing pencils is zero, it is necessary that M' charge be induced alon:; the pened that g 25 lr will produce an electric field jus, equal and opposite to that which had existed }s 20 l* previously. The induced charge will vary fy ' ([ nuan aa,c Qg linearly along the pencil and wdl be pro-portional to the distance traseled by the
)
"y- - - -- -9 5 . is.4 o, tip. The charge density, and conse-l _ .
Jo o5 40 quently the electric field, at the tip will
'S f 4o j
be proportional to the spacing of the sr electrodes Now if a new term, F, is defined as the l
- 0 o o5 to 45 20 23 30 35 40 45 50 so 60 6s 70 oservoltage factor twc N wCRCSE0oNoS FCR Nmio5 OR -$ T v.t" -- #
us (12) Fig. 8. Calculated time las curves for rod-rod saps for various applied voluse waveshapes and equations 11 and 12 are inserted rod-rod gap in response to a rectangular velocities of the channel tips, as will be was e and it is assumed that this quantity explained in more detail later, are pro. ,_ x/s . jc , ,, ) (g37 is proportional to the gap ten.;th, then portional to the electrode spacings. V-xa u,,- mx in kv (8) In this same nomenclature, equa-the physical considerations invol ed. . If t.he ditTerence of the applied waveforms Suppose as premised earlier, that in their
** ' Y'*" " $ I ""
changed to the following: is taken into consideration the factor m development the corona streamers de-
-t[uys corresponds approximately to the average posit a charge density in the interelectrode Jr I sparkover value in Fig. 5 in kv ;>cr em space, such that at the instant of chan. 2 " ~ ** _m( 1/x _
Equation 5 can be rewritten as nelinitiation the electric gradiedt between y dx ur the electrodes is essentially constant =~sKm y1 (14)
-=-k --m (O and equal to the value m. Furthermore, Jt x if it is assumed that this space charge is w hich confirms the previous statement which merely states that the selocity is relatively immobile, then as the are that the velocity is proportional to gap proportional to the excess of the average plasma develops within this space, lengt h.
gradient across the unbrid;ed portion of the gap over the critical sparkmer gra-dient, Cererally the factor k will vary t2 with different gap lengths; the negative 5 o voA3E sign is inserted for analytical purposes so it may be recogni7ed that the un- $ -s- .
.cy's bridcd gap decrea3cs when the quantity i to - l
" " - " " ~ ' ,
with the parenthesis is positise Further ** 'f transformation of equation 9 is possible f o, . 0 to the following: $ \ l g os - k"! sg,, s
-(x) t
-g !
s
% t 7 o' *
. u_n o m s s a
5 os - j 3 l The right hand side is thus te !uced to a a per-unit g'ip length basis. j 33 b Rusck's equation 7 which is applicable ,e I )
.to rectangular voltage waves applied to j ,, _
tt j 1440 cm gaps shows that the time lag Fig. 9. Determination of d is inde;)endent of the gap length. The the miation of channel 7 8 - kl
$il 4
nlocity, current, and length y vnoarv , work of McAuley" with 1.5X40 impulse of unabridged gap with time g waves on gaps up to 100 inches when re- or f , an applied 0.5 x5miec g plotted in Fig. 7 shows a similar inde-pendence of gap !cngth. If,in equation 0, ,', '8[,[* d ne a - o,, cupau r , hj ' I 4 - Ks y I (11) dotted line afs curves for an applied surge with crest over- $
" 1 !
then equation 10 is also independent of eoitage factors of 1.25 and o io I2 'o so ao gap length. This simply means that the 1.10, respectively r m w menostc acs
, \
Ocioun 1DGl Wagner, Hileman-%chanism of Breakdirrn of IAboratory Gaps 61L 90004115 i
/ l
*C voltage ratio of 1.0 at which time the dominant characteristic. M ost of the l 6
] j channel begins its travel across the gap. difference in time lags for short times is I
.' I Therefore, the time denoted by the dis. due to the differences in dead times.
soo _ _ _ C u .
, . tance c is actually a " dead time"; that is, Consider now the critical witage for a
' 'l during this time the voltage across the 0.5X5.0-usec wave. Acenrding to these
,,y ,
, ! , gap is not sufficient to initiate a channel, calculations and theory, at an oservolt- ]
; so cul The total time. T, for the chanr.el to com. age facter of 1.25, T4S usec. It was
; f .
g soo *' '
, plete its passage of the gap is 3.S ssec. noted in the calculations that for an i
- 8 l l Also, the velocity of the channel with overvoltage factor of 1.10 the gap did not
, 5 ! -[ respect to time is shown in Fig. 9. The spark over but channels were initiated.
I
~M significance of this curve is must easily This is illustrated by the dotted curve of
] '.l- i
; } !
visualized by rewriting equation 14 as Fig. O which shows that x/s starts to de. l crease when the overvoltage ratio ex.
'oo -
A l {
"\ x/s /
act., or w io ("
' di w s e tail causes a rap.d i decrease of volt.
e . I
' '
- Therefore, from Fig. 9, the velocity for age, the r/s curve reaches a minimum tw[ oms [ves value, and then tres to its ongmal value any specine time is the distance a divided Fig.10. Calculated time.las curves for a rod- by distance b multiplied by the constant of unity, rod gap with an applied rectangular voltage 4Km. The current curve of Fig. O is wave es determined from equation 16 com. LutRPREtatios car Test DATA discussed in a later section.
pared with Rusck's" test data indicated by In Fig. S the time lag curves for several Rusck's data are convenient for testing
'! the points the validity of the relations presented waveshapes for Km=0 5 are presented.
X-Positive polaney The overvoltage factor is plotted for other here because he attempted to obtain a
; O-Nesstive polenty than the rectangular wave as defined by rectangular applied vcltage wave, it equation 12 except that u, is the crest rose to crest in about 0 3 usec and was
[ voltage. At sparkover times when T is flat thereafter with the absence of oscilla-
} ArrucATtos To PART:ct1.AR less Aan the front of the wave, the over- tions. Fig.10 shows his test points for i
WAv0 SHAPES voltage fa@r plotted is the crest voltace rod-rod gaps of from 10 to 70 cm. There With !* known as a function of time, actually obtained across the ga p. In was no appreciable diiTern.ce between equatien 13 permits solution of the other words, there curves are constructed pmitis e and negative pclarity. By choos. determination of the diminution of x and pk;tted in the same mar.ncr as normal ing Km = a40 and m = 6.15 kv per cm, with time. Solutions will be obtained for time-lag curves. As noted, the time axis the curves represent the computed results two waveshapes. can be changed easily for any other value for a rectangular wave. For times lonrer f of Km. than I u ec s the agreement is very good, Rectanpdar IVare it may be saen in Fig. S that the critical about as good as Rusck obtained with his l expretsion. But below 1 usee Rusck's l With V constant voltages va inversely with the wave tail but are independent of the wave. relation shows a better agreement with
! , y, I
- fmnt. For example, the critical voltage tests.
l Km dt - db- Fig. 11 is a reproduction from the 8 a . V-x/s \s/ for a 1.5X40wsee wave is about 1.05 and f y _ ,f, for a 1.5X 10 cec wave is about 1.13. paper by Akopian, Larionov, and Torosian Kmt - -(1 - x/s) + P !n (15) lionever, the critical voltages for a 1.5X showing the time lag curves for several 40- and a 0.5X40 asec waves are equal. different applied voltage waves for a CO-and for complete sparkover, the time lag As expected, with small values of time em rod. rod gap. The relations are the Ti8 the reverse is true; that is, the front is the same as used here for which h or Ks was Km T- -l + V in V-1 (16) , - 900041 L6
! l l- l i I ! I
'l 1400 - -
The value of T is plotted in Fig. S for lr l l l 1 i l l l Km equal to 0.5 but the curve is applica. ] l l l j ; iT ble to any value of Km. This value of , :20e 7, ; ; ;^ l Km was used as it corresponds to a valac -
~
that, as will be shown shortly, fits the l i ki ,
~I- '
' l l
' l t 1
icoo -
^
! observed test data of Rusch and of I 3 ! [3' _-
t Akopian, Larionov, and Torosian. Fig. 11. Time.las curves for 800 l l}{ ; , , g Linearly Rising and Falling IVares
- 125'cm 'od *'* d 9 *
- c *I*
culated by equation 5 com. lg' sco __:I' ~~~j i l L. ~ ~ M ~~~ ~ The time lag, T, for a specific over-voltage factor, V, for other shapes of the pared with test data for applied positive ,3 clarity volt. ~S'
~
!l i l j ;
[ rm! !Q' , l
}. applied voltage is most cons eniently age waveshapes ,;s i!!ustreted 'co J2 , , 7 .[',- ,
]
determined by using a step.by step solu. In inset" i g hbl , P! i ' tion of equation 14. One such solution 200 ' ' ~ 1 --Sta nda rd 1.5 X 40wsec i,! i for a 0.5X 5 0pec surge, w hme crest over' wave, ta = 2 usec ' q voltage factor is 1.25, is shown in Fig. 9. 2- to 2.9 usec, <:bse ratio !
' ' __ w o
i The quantity x/s remains at 10 per unit us/ui - 0 6 0 2 4 6 e io i2 i4 until the applied voltage exceeds an ovei 3-ts =1.8 usec, us 'u n= 0 63 DHME IN WRCSCcNDS
, 612 l1*cgne r, Hikman .Wchankm of Brcakdown of Laboratory Gaps Octour u lunt s l
e > t gn , flg. 12. Voltages , qg 9 ; p across and currents e -. through a 6-foot ; cool!' '. : A vertical red rod gap 5 i' ,' { 8"
}
~
for positive polarity } #- ~ t ! j applied surges. The p,,n,
-3 J_is o * - - - - - - -!
T sur3e generator d ,en , i a charging voltage for o - 1 . _ .. ._%.
$ 2tr
)
criticsI sparkover is g.-- , j , ; k iso g ' ' 8 aoo , s7o - += OL 13 0 i s. f'
- T7- '
iso ~.- - 3c00h*'I j 'i ' g. y2000 . o f,
- 2t2 '9 o_ydus a 6 i l(000 h p M [)
- ~ f'-
dj j i
.f ,[ -
800 5 8 [$* aI g t #** x 6Fr g ; , i s : 0 so l h n( $. _ { 600 - 1,," %. % Ito 3 8 2 s 4 g - t6ur
- MicRcScc0Nos 3
g h ; ., cuanomo Fig. 13. Typical oscilfograrns of voltages f400-v S ocrace tso across and currents through two parallel 6. foot tod rod gaps sepeated 18 feet. Charging zoo . voltage is 200 volts A-Veltage acress gaps B-Cunent through gap which sparked over 7 i 2 s i t,ur a wcnestcovn s s 7 C-Currer.: throvsh gap which did net spark
+
over when ether gap sparked over i
- 11. The vah e of K is then 11/125 or floor. The voltage across the gap was 0.0SS. From the 1.5X40-asec curve the measured with a 21,000-ohm com-pensated voltage divider and the gap of the inductance and the capacitance.
value of m was estimated as 720/125or 5.76 kv per em. The factor Km is then current was measured simultaneously by Its effect has been ignored by estimating the current, when necessary, as the aver-0.507 which may be compared with 0.46 means of a shunt located about midway in the lower rod. Successively higher volt- age current during this period. used in computing the curves in Fig.13. An average value of 0 5 might very well nes wcre applied to the gap by increasing In Fig.12 a number of curves of voltage have been used in both Fig.10 and Fig. the charging voltage of the generator. and current for ditTerent charging vol-1L The other two curves of Fig.11 in- The charging voltage is an arbitrary tages are plotted. Contrary to what dicated the degree of agreement obtain- number depending upon the a.c voltage might be expected from the theory just able with widely different waveshapes. In applied to the low. voltage winding of the presented, current does not begin to flow their computation it is presumed that transformer, which supplies the voltage at just the instant that the critical break-that is subsequently rectified to charge down voltage is exceeded. The time de-Akopian, Larionov, and Torosian used the experimentally obsersed potentials the capacitors of the generator, liow- lay at which current is initiated is longer, ever, while arbitrary, it is a quantity the smaller the excess voltage overcritical. directly across the electrodes and there-fore took into account any internal proportional to the voltage to which the The delay is made up of two components, generator is charged, prior to being dis- first, the period of waiting until a free drop that may have existed in the surge 4 charged into the test circuit. For the electron enters the overstressed electrical l genera tor, i ! critical voltage of the 6-foot gap the zones at the two electrodes when the charging voltage was OS volts. The re- corona streamers that form the space Experiments by the Authors charge are released, and second. the time sultant waveshape of the critical voltage The authors undertook rneasurerrer.t of is shown in Fig.12. Increasing the charg- required for the development of the space the current in long gaps under sparkover ing voltage resulted in drawing more charge and conditions propitious for the a conditions in order to verify some of the current from the generator during the dis- formation of the channels from the elec-i discussed ctmeepts and also to study the char;e process and this current drawn trodes. factors affecting the current variations, through the resistance of the surge gen. It was impossible in the open condi. ;' l because this is the most important crator resulted in considerable distortion tions of the laboratory to obtain a clean- " cut oscillogram of current supplying the p variable to the transmission engineers. of the voltage across the gap. Figs.13(A) A surge geatrator consisting of 301/4- and (U) are typical oscillograms of the space charge just under critical voltage gi as obtained by Park and Cones + and g microfarad capacitors was used. Other voltage and the current. The inductance constants of the circuit are shown in the of the surge ger,erator generally does not Degnan." Apparently the high free t-electron concentration caused triggering y, insert of Fig.12. In one series of tests play an important role. The oscillation a vertical G-foot 1/2 by 1/2-inch rod-rod in the e trent and voltage following com- of the gap on the rising portion of the volt- j , gap was used. The tip of the lower gap pletion of the passage of the gap by the age wave and prevented the sharp rise 'i was about 6 feet above the laboratory are plasma is caused by the interplay and exponential decay of the current. l 613 : OcTonER 1DG1. \\'agner, flilen:an-Mecitanisrrs of Breakdm of Laboratory Gtps L -. 90004117
I do Fig. 14 (left). Rafation be- : - - - 7 tween the charge Ied into th3
'cco - I pla.ma channel and geplength ;g ,
for roded gaps { p ,ng, ;p 8 "C
!ac l - -
?
3 B o 60o - h l . g is - 9 FCOT GAP ~ E 400 t-I o i e "
, o ave
- AGE v& LUES g
. le 6 Foot GA -
s
* ~ ~
j Fig.15 (right). Emperimental $a3 . _ time.tas curves for 3., 6., and C 4 e 3 . , a 9 footrod rod gaps for applied l w a c,m m n positive pofarity voltages as o io zo ao ao so so ro eo
! illustrated W E W E HCsE;cNOS i.
The results obtained did suggest that if In Fig. 9, the vebrity was computed both would spark m er A current shunt the diwharge had been delayed a crest of for a 0 5X5#see wave and an overvoltage was placed in the grounded electrode of about 25 amperes would have been ob. factor of 125. .\pplying the factor Kr gap A only. The upper curse of Fig. l tained. 3.2 to this vehicity curve gives the current 16 shows a replot of the current w ben ) curve indicated. This should be com. gap A sparked over and the lower curve CtunT-Tiste RELATION pared with the current curve in Fig.12 when gap B sparked over as a voltage of Saxe and Meck'$ concluded that the for Cl% 130. The comparison, though 2000* c of critical was applied to both gaps.
! current "is proportional to the vehicity of not perfect, shows a general agreement in Fig 13 shows the osedhr.; rams applicable the ka ler stroke " For the present this nature. to this case. Initially both gaps carried
; relanon will be accepted and it will be . , current equally but as the channels de.
awumed that the instantaneous value of Veh' ped one traveled slightly faster and the iura nt, i , is proportional to the t With a given surge gntrator setting hence drew more current. It did so at the inst antanmus seb. city. Thus, havin; no adjustments made to maintain expense of the other which then did not a particular w aveshape, the time lag have quite enough current to maintain l 9 , _ g, '!! 1jg, curves, according to the theory ; re3cnted a corresponding vehicity. Furthermore, j e here, should be independent of :;ap length. the first one decreased the unbridged gap I I i T.ne ne;ative sign is introduced inause This hanens because as the gap is and tended to travel even more rapidtv - t, doubled, then w.it h the same overvoltace than the other. The etTect was cumula-as the unbo.dged gap becomes smaller
; factor, the surge generator volt a ge, t.ne and the one to spark over robbed the si.:n of dx/dt must be negatise and it the velocity of the channel, and the cur- more and more of the current. Tln.s effect is deu.rable to consider the current as a rent and the voltage drop are doubled ated was pronounced only after the differences
, pmitn e quantitv.
; : the same time lag should result. There -
m velocities and the lengths of the un. if this relation is. valid, then upon fure, if these relations are correct, the e integrating both u. des one arrives at the . bn.dged gaps became great. While the I g; time lag curves plotted against over. phenomenon is essentially a resistive one, voltage factors for 3,6, and 9-foot gaps largely dependent on feeding an appro-l f,r ,g, , _ g fl dx should form a continuous curve. This is priate amount of energy into the channel demonstrated to be the case in Fig 15. to raise the temperature to those of an C*U % arc, undoubtedly charges also rush into PFoGRESS oF CH Nt.S This states that for any particular value of the channels as they progress and the s, O <hould be constant. .The area under Two vertical 6-foot rod-rod gaps were Selds between the approaching tips in-
. set up IS feet apart so as not to influence crease. But upon contract such charres any one of the curves of current in Fig.15 -
, . each other electra. tan"~ M. hen prop. rush tow ard och odier from the opposing to the ir stant of short circuit ts the total '.
, erly adjusted, on apph. cation of the surge channels through the completed paths.
charce r.e d into the channel for a 6-foot potential, one, the other, or sometimes Only an inappreciable amount of tla.s rod rod gap. Tlie range of values thus obtamed from Fit 15 is shwn bv a bar ' m Fic 14. SimiIar results obtained for a 3 fmt and a 9. foot rod rod gap are also .. noo plot t ed. The slope of this curve gives a 3,, g calue of K, equal to 3 2 microcoulombs m c vE n N o se ct
+ 600 LE M h a r7R --
I er em or amperes per cm per sec. ,This , a
!ini anty serves to confirm the propor. 3 ,_ L, tionahty exprened by equation 18.
^
E '" 5Fr ser Saxe and Meek presented a similar curs e obtained with a positive rod.to-plate 5
"C -
[.e- hCJ'"7 - i gap for gap lengths of 8 to 55 4 cm u hich sh .ned a re markable linear relation for
~
# 0[n" Fig. 16. Current in gep A n hwh the slope n as 0 W microcoulomb den ;ers A and B are im.
per c m. It also kars out the general puised simuhaneously. Gars. Oo [- c3 ,o s zo 23 [a 33 40 nature of the phenomenon. ing voltage is 200 volts T YE % M M E N s i an wawr, ni: man-wi: ant >.n of nrada o.r Labora,'ory Gaps Ocionen n 1 900N ' 13
N r- 7; ,ppn Fig.17. Channel formation in f' ,.y
~~m-m,a,;k'N2T[h
[* q.c ; - 3i*
.m- , p.
an unbidged gap when the oth er gap of two S-foot *o - m, 7 _ c-4
- q g
patalle! l rod rod gaps sparks N 1 - C.__ ( y over C # ' 'S J 2M' N d i E N IO
* ' M 800- l d
e I
., 2'6 a._.* su / -' sr.*u ~ ~Aq char;e is olacrsed esterr,.dly as evidenced ;;ener ator. Neither the K nor the K, o, , , , ; , ,
by the absente of a negathe current in constants applicable to rod-rod gaps are the unbrid.;ad gap following sparkover applicable to such a gap. The multiplic- r>oo itv of parallel channels apparently affects ev of the other gap Further evidence of ,, N,3,6 this pro.;tew of the channels is otTered the tields near the tips of the advancing [ zoo I by the still photograph shown in Fig.17 channels and retards them as compared j %s taba of both ;ips sunultaneously Note with the few chanmk in the simple rod- g eco r 7 I so the euem to whish the thannels in the rod gap. The K eon:t ant was determined 7
.;a p that did n"t sparWr hate ad- only apprminiately an I was found to be {" 4 col vanced. about 0 05 to 0 00, which is smaller than that for rmi rod gaps No further work o Rtsc,Rtm Gae's was done on this gap at this time. .j , %c$cm [e, A 72 int h-diameter rmg made of 2 Pter-Ptec C AP Fig.18. Vohages across and cunents through im h ;ipe w as m'iunted di inchc3 above a a W. inch ring ring gap A in<h ring, the latter was hcated 6 feet An enlared form of the ring gap wa3 above the laboratory floor. Sur.;es of set up, pritnarily to simu! ate a long paral-mitive polarity were apt lied with sub- tel pipe gap which would have teen im- a ce These are formed becau-e of the stantially the 3ame surge ;;enerator con- pu3sible because of the restricted space of time required for the fonnation of the aants as >hown in Fig 12. In Fi't 18 the laboratory Three 1mch aluminum space char.;e- The channel eurrent crms t he ' i ol qr2m tram "t .s 71tas;es and cur pipes each 12 feet lon.; were arran.;ed .m v .- d!y and if there were no induc-rt nts are ph.ttel T be critical sparkoser trianpalar conh.;urati n ab'ut i feet fram tance in the sur.;e generator circuit would voltage occurred with a charging voltage the laboratory dour and a sirjatar ,et was increase almost vertically. Simultane.
of Ilt The character of the predis arranged 6 feet above it. Fig 19 shows ou>ly a correspond.mg drop .m gap vnlt-charge ennents is quite duTerent from corresponding volta ~e ' and current traces. age occurs For example, consider the and of much greater magnitude than for The currents were even targer than for t!le dischar,;c for a chargin.; voltage of 200, the 6 foot rod gaps. A very lar;;e drnp ring rin.; ;ap. A very ;)ronounced pip TM Mmm mW dW h occurs throu;h the remtance of the sur;c occurs at the begmning of the voltage surge generator produces an internal drop of 900 x 1.000 or 900,000 volts The Fig. 19 (leh). char.;e draw n from the surce generat"r 200n-Vohages across capacitor produce 3 an additional but
considerably smaller drop w hich is directly
.g.
4 [ - and c urrents through a 6-Ioot calculaNe Pecau*e of the distortions i
- 6FT pipe pipe gap
*5 2CC- : , l
' 12 F T :200 N
> f !
800- r ev=88 r 800 3pr
' - ! N2 FT
!60 -
cv so 40C iO4 , 3 f M is0 10 0 ol t 0 0 2 4 6 8 10 i2 '4 7
, p
; om _. _ __ _ _-_ _ -
2000 - ga ~ N .. ,s -l 9 bCC" 600 i q !?OO -
\ 16 0
!h (200
- 90 d i !
e- - 00 3 9C0 j 7 goo - i > # 4 l 'l D k 10 4 ee - Fig. 20 (right), d 3 so -l eco- { Vokages across and currents ' I
~
a
^
2 4 6 a '5 '2 4 through a 3-foot o i 2 F4 s 6 ' TwN wcocsc.cN M pipe. pipe gap r w % v ocen cNas
!!'agrict. H'le enars-- M,cha rri3rn of Breakdezn of Labomtory Gaps 615 Oct onut 1961 90004117 -
/ in the waves it was difficult to line up the 4
,. function of the phenomenon occurring reference points precisely but it can be $ , a l within the gap, as the regulation of the assumed that the rapid rise of the cur. [s -
*- 4 circuit is sumciently stiff that the cur-rent trace should occur simultaneously ;x rents required by the space charge do not with the abmpt drop in gap voltage. (2 produce much drop in the external cir.
Fig. 20 shows a corresponding group of
- cuit. While Akopian, Larionov, and curves for the same gap set for a spacing ;' -
Torosian used a somewhat larger ex. of 3 feet. Esen larger cmrents result for ternal resistance it can be anumed that a particular overvoltage factor. The {-c for the length vi gaps and for the rates of o . . , y go channel currents were mainly htmted by sont-cscar evaao.r m es tats rise of voltage they used, there was very the surge generator's ability to deliver little lag between the voltage and the fig. 21. Energy fed into a 6-foot rod od gap establishment of the space charge. This higher currents. The high currents are g j attained with only modest increases in the explains why in their analysis of the electrode voltages. With luwer internal ,r. RKJus (22) time to breakdown the phenomenon re=is t ance even higher currents should ec.uld be described in terms of the develop. j result. which demenstrates that H, .ts pro- rnent of the highly conducting channels I Nu detailed analysis was made of this portional to I,, and s. Since in equa. tion 19 it was assumed that the current type of gap at t!..is time, but just as the For km*er *-aps, such as the 200 inch tharacteristics of the rod rod gaps are is prcportional to the velocity, then the . rod. rod of !!a,;enguth, Rr hlfs- and h.ueanty of the energv-current relation useful for studying the nature of the ' Degnan, the time of space charge forma-tends fu le support to this assumption stroke proper, and will tx considered in
, this connection in a companion paper, -
time to breakdown. It remains to be the characteristics of the large parallel . ascertained whether the time lag curves pipe gap will be discussed further in a General D.iscussion g
; paper concermng the performance of the can he cumputed in a mam.er sirtilar The authors have kept their evnjectures
, transmission line tower. Tlie K con. . . . to that emploved by Akopian, Larionov' c neernin; gasevus c!ectromes to a nuru- - -
stant was found to be apprcnimately and , Fores.ian and amph., he d in this paper. mum and have connned themselves largely O'n'd to O Sr5' The faegning staten.ents may be not to the esternal manifestations of the E .r ec,v Ft o INTo Tac D:senacc phenomenon. The physical appearance comp!ctely s lid for conditi"ns near the of the coruna dixharge alone is ample end of travel of the tham:el. Ifere the The.: stantaneous values of current and s ei.. city attains very high v. dues und the endence that the gasaus electron es votage from Fig.12 were multi;1ied and phenomenon is quite different for poso. space charge may not be able to deselop
, inte;; rated to give the energy fed into the sut'ie:ently to keep pace with the values i tive and negative polantv, but externally -
I d.echarge during the breakdown process. ' correspr nding to the reduced unbridged they differ only .m degree. N.evertheless, I The results of this computation are WP-the audwrs wish to conunent upon some plotted as circles in Fi*'. 21 against the As the good conducting channel ad. aspects of the discharge, short. circuit current, I,,, of the surge gen- - vances thrcugh the relatively immobile It was menti ned in the discussion of eratur. Four additional points obtained the epate charge that os er a considerable sif ace charge, as has been mentioned pre. S months previously, also un a 6. foot rod-rod gap, are plotted by crosses. While rage gaps a te r a m age nou@, %es are inM upn eas
*lectric gradient determines the long. pencil of are plasma. From the este. l more than one channel is involved some .
mates of the current m the channel and time applied voltage at which sparkover l portion of the time, the straight line in- from experiments, such as performed by dicates the value of are energy required occurs and it further appears that the g ent is constant along the center line Ifigham and Meeknon the charactenstics to rise the temperature of a pencil of the of rapidly developed arcs, it can be cun-gaps to are temperature as 105X10_8 of the 6ap. The work of Akooian'
- Larionov, and Torosian also indicated cluded that the diameter of the arc joules per ampere per em. plasma is approximately 2 mm and that this as the channel began to develop
, The energy required to des elop the the are is a relatively good conductor. A w en % mnesp<mding to the particular
, are should he linearly proportional to the short circuit current and the length of the g p was exceeded. By "lang time" in high char.;e q mduced in the head of the
. this connection they implied a time of the < hannel that is conducive to the develop-
. mt
. erestmg to contemplate gap. It is order of 100 rec. With longer times it ment of a high gradient laterally as we,d w hether this is consi< tent with the relation as ahead of it. Tlus c har;;e and grament, i' quite c :nceivable that other factors that the total charge fed into the produc- m tur 1, give rise to copious corona dis.
nught enter which would alter the nature
! tion of the arc, such as plotted in Fig.14, dages. De head ha end im@
is proportional to gap length only. As-I O e space a rge. Thus, for a sus. tained and continuous potential, as in the process which m,ght i be termed a suming that a rectangular voltage wave counter corona discharge that takes d-c corona. the charge distribution might l' is applied to the gap and that II' is place withm the cn; mal space char;;e. be quite dderent. the total energy supplied to the "ap, .. Conditwns conducive to the development It was fourd by Park and Cones that i Omn of such charges are present even thcugh when a 0.07 X 100uee wave was im;>ressed j !!% VQ (:o) across their sphere plate gap, set for a the field ;;radient is not uniform as pre-i' letgth of between 20and 50cm.the current mised by the foregoing simpli'ied assump-If R is the series resistance, then at short that supplied the space charge roce to tions. I circu t crest sm rapidly and then decay ed to More than une channel can form si- l h"" ' "' " "S* 'o it may be said multanenusly, but as they progess in II% RI ,Q (21) that the space c.harge is substantially paralk! one will advance somewhat far-and substituting Q from equation 19 established in 0.1 rsec. This period is a ther and ter.ds to thield the otherselectro-
, 616 1f *agner, Ili!cman-Ifeci:cnism of Brmkdun of Idaratory Gaps OcTohnR 1901 90004120
statically and thus reduce the Eeld in gap the electric gradient in the pp be- contra: ting characteristics lead to an advance of them. By this process the tu een the electrodes is approximately uni- explanatioa of the steps in the lightning advwe of the others is retarded and this forrn. The avera r,e critical gradients vary stroke. eFect bee tres prege Ae. For gap between akut 5f.oo and 10,000 vults per The fact th tt the mapitude cf the conhrat Lns thit apprw h two ge e em dependi: , up n ;ap cer.63rd n anJ current > feedin; the inimi part charge rnetric lines paralk1 to each other such as polarity. Wht n the critical avera;e are quite sn:a!!, in contpa ison with the ! formed by two long pmilel pipes, this gradient is exceeded a channelis initiated currents that cecur during the plasma e:Tect should not be as dominant as for which usually starts from the anode. In channel forming phase and with the a sinde rod.rnd gap. The te3ts rnade the case of a rod-plate op with the rod short circuit cunents permitted by the with two G-foot rnd rod gaps set 6 feet positive the char.nel develops fr tbc constants of the surge generators, was apart as shown by the insert of Fig.12, entire length of the gap without the de. appretiated quite early. As shown in showed that in some edes both gaps velopment of a pla ma chanr.el from the the ci>mpanion paper, the currents occur-sparked over simultaneously which in- plate. But with the rod negative a rin; during the steps of the lightning dicates that for separations greater than plasma channel is first initiated from the stroke are also small in comparison with the pp length the shielding etTect is not plate, and after progressing about half- the currents in the return stroke. The l verv great. Tests with smaller separa- way across the gap it is met by a m- phenomenon appears to be essentia!!y l tiim were not made. Allibone' showed rapidly moving chanr.el, w hich starteu at a thermal one; sumcient encrgy must be that e,en for two paral!el plates two a later time, fr. rn the rod For a rod. injected into the pp in order to rai e dominant are paths can fur n simultane. rud gap, the plasma channel also starts a thin cylinder of air to are tempera-ously from the anode and is met in mid. gap tures. .
)
While more information is available by a later initiated channel from the con ( erning the propagatior of channels cathode. The drop in the plasma chan-frma the anode than fror, the cathode, nels is so small that it is considered References sini c the pr >cus of cha .nel formation is nedicible with respect to the applied 1. A New tri m en N rua C u cetaries or es entjally a thermal procen it is ex. voltages erancerned in th.is pheni'mesn. rue lm ii r N m. Purun m . o, Te e m:w l pected that the veluty of propagation of For rod-rod pps. the heads of the two ((N1 'i[E*O,js.,$"y'N",'DC.' ) the channels frotr the cathode should be channels apprciach uch other with a m n 12 u w i; of the same ordt of ma;nitude. velocity that is proportional to the excess 2. A New Arenoacu ro rus C A{. COL ATION OF l
. . ru tt LICn T N D.O PE RKUM 4 NcW OF Ta A h nt !WO N ( i Of die terminal voltage over the crtticut LiN , -!!, C. F. Wa g :er, A. R. IIHema n. IS L, ' '
g spirkT er voltage for the instantantcu.i d M U" 19 M N D WL g. 3 ^ N " " ^ "" " o ' c's t o rur cuct1.ariox o, value of the unbridged gap, and inverstly rH R L1G u ? % I N G PFuPORM ANc5 OF T a 4 N WihlOM Upen application of an impulse voltage, pmportional to the terqt,a of the un- LM -HI A Sm runo M a runr-Siscu s to of such value as not to cause sparkover, bridged gap. The channels grow with a 7'9$ Q'yl'fh^;o"
" 8 " " ' * * ' '64 l
a nonuniform field gap, of the propor- relatively small initial velocity which is t .rne Linu miso stem. c. F. Wagner, A. R. I ' tions frequently encountered in engineer- accelerated as the unbrid;ed pp de- nuem.m th t , voi 77. June 19w pp 229-42. ) crease s. By using these s elocity relations, 5. DETr umannu ur W4ve Feost or Ltc u r- 3 ing work, the f. eld at tirat corresIands m o StaOKF CURu%f 3 FRUM Fin o M s Asens. I g!" "I** }^ 4 { f g' r g pps can be to that which would be expeded f;orn m rs. c. F. wa gu.r. thJ , vet 79. oct 1960 l the cons entional electr matic solution. computed for anv applied voltage arras op 581 4
.the ;ap. Expreuing the applied voltage 6 Ta= L cu r so.o sinnan-rt. c. r. waca.c ,
The 6 elds in the vicinit.y of the electrodes A. R. fluerran. thJ rseepp 60?-42 of thu msael, in terms of the critical sparkover voltaze e, , mar exceed the critical Geld momentardv 7. stacs T,ot. r a c e BarAKUO%N OF Ain 1N a f r a rectanedar wave, results can bc I xos n mN F,,w y, n , p. , k, H. N. cones. but when this tiel! is eveeded and a free C* % elettron appears in the re.;i'n of the reduced to a per-unit basis that is in- {ag" 'gf W3d"aba D l over<treccJ tkld. an electron avahnce is trigsred that de,elopsinto a spac e thar;e. dependent of the length of the gap' There characteruttes can be completely
- s. stu m a c e n n e h rs tu nica voorms sr.n ov, . J. s:erun, J. J. Torok. rv u.c
)l 26, 1929 pp For rod rod gaps the space charge de- described by two parameters. g[ Eau hushrd, Pa , vol velops from both electrodes but for rod. The channel current is proportional to 9 Tu, u,cumu or ma Lono seau. T. E.
plate gaps from the rod only. The flow the velocity of propagation of its head, * * " ' /*"t rnmutma of Electruat Ensi. peers. London, Fngland, vol 97, l9%. pp 313-2 L and therefore can be determined in terms of the charge into the intenening gap is at .
- 10. DintrcTaic Puesovr y 4 Ar liten voo rrors, i a rate of ab. ut 01.01 c which corre3 pend 3 of the inantaneous velocities discu%ed B.L.Gon W ,F.5 Ed war d s. F. R. Perry. 16s t ,
apprNnt itely to the electron drift; so p res i~t: A For sadl oventltages the """"PP**~ T S r '"m " o " that for a 10 cm pp the char ~e ' has dif. chamm arre nts are usua!!y cona e ".ana "" Dixu m c. 'v". PKm' ' ' e" "m" 5 m', ..,.'n .'. r t l i upwa- but for high ovenoltages and wu or mence. Tnwa sees sumn., f ' fu<cd throu.;h the entire pp in about 0.3 sutkiendy high resistances between the
- EM*"* U58" '* W * "
fee and for a 20 inch gap in 9 ecc. 3" t T The current feeding the space charge api lied voltage ar.d the pp, the current l,2 p, "" $"l',"&"['gp "' %' '[Q 3 i rises sery rapidly and decreases somewhat is of a stepped character. /w ~t insoon.n of Eintoal Enmmt W <
.. 107. p C, tG -)
akr.g an exponentian cun e so that a The form of the current wave feedin"* 13 Snry Cvet.e aw f vert.ss serenavra ano i a substantial portion of the space charge is 9 the imtial space charge when a rectan- Luna c., 8,.cm J. n. Hagenavn A. F. established in slightly less time than these Ro ds. W. J. Degu n. AW Da n mm,s s. gular wave is applied to a ga E and the form pt. !!! (Patrue A ,a,os m e ,sg) Syns< m s t , vol . 1, , i Values. of the current flowit.g during the develop- 19M. pp. 4!GeA j A certain critical average gradient exists ment of the channel is quite opposite; to ronuinou crursrs.w ru ouant.ow, J I #"" for gaps which wdl produce ultimate the former decreases somewhat as a nega- O,'l'n'.'%ol/,; } ]I[ " "'"' ! sparkover of the gap with prolonged tive expcnential with time, and the latter ts M ren - o, % car Nica nv a se==r application of the voltas;e There is increases somewhat as a positis e ex- D'Sc"^
- c r a i
- Aia at A r = osra m 'ic Pa nwa s, some evidence to indicate that when the ponential with time. It is shown in the UJ '. $"Ne1.N'3.1((p'.'M.Nl" ""
spate char;e is fully de.cloped acrus the companiun paper in this issue that these i s. turtta w co cycu sruscrn o, Aie, OCTuuLR 1901 M*awr, !!ileman-Mechanism of Brmkdn.cn of Libomtory Gaps 617 l 9a0A412' !
/
e i P. L. Betlaschi, W. L. Teague. AIEE Trew 9t lit (Pever Apparaf at sad Systems), vol. 74, 22. Errsc7 or Non ST ANDAsD Setos VoLTAoss wt.1955. pp 85&-72. om IkscLa nox, Sune Rusck. Bullina no 40s, h d(I5o4s, vot 53,1934, pp.1GaH5.
- 17. Iurt Las Cn Amacranistics or L4aos C Ars. 20. Baue Peneusses or C Ansors FLacisowies
;c[
A. A. Corer, A. M. Zalesk y, B. M. Ria bov. B us' fimk), L. B. Loe b. t'aivermey of Califve rus 23 FL A%uGYes CH AR ACTE 81sr4Cs Or INWL ArtuN, ise no.141, CICRE, Pans. France,1948- Press, Berkeley and Los Angeles, Cald.,1955. P. H. Mc Auler. Eh.fisc luminal. July 19'I8. pp. lt If 18 Rsmar on tna Wc.az or rue Srvov Cou' 21. Ou turt'tr.s Disen Asos Votracas Acausa
, uittas No. 8-Ovss VoLTacas Awo Lmstmuo. Hacu Votracs InscLArion As ReLarso to rus 24. Vol.t ro s Op Ausss rs in Lono CAskoeg K. Berger. E=lleHe no. JN. CICRE,1956. Suars or tus Vottacs Wavs. A. A. Akopian, Sr4:n Carorts. J. B. Ifigharn, J. M. M eelo V. P. Larionov, A. S. Torosian. Sudeba no. 4ff, Pror,rdings. Physical Sociciy, London, Frig!.nd,
- 19. Errsc7s or CosoNA on Tsavauho Wavss. pt. 9, vol. 63. Sept. MO, pp 033 4 48.
C. F. Wa gner, B. L.1.tord. 4/EE Trassertions, CIC R E,1954.
+
- voltage to its peak value, umi, t he capacitive sometimes iu, may be observed. While DAIS CU S SIO ft chargingcurrect with peak valuei t m. flows, they are still present the final rise of current At the moment in, a current impulse which up to the breakdown at to, follows. ,
i H. Baatz and A. Fischer (Studiengesell* may be called current of impulse curona The inception voltage u, of impulse schaft f0r Hochspannungsanlagen e. V. , appears. Its peak value is in. At in, the corona increases slightly with increasmg Nellengen Ober Esslirigen, Germany): The voltage at the gap is ut. This value is the peak value of impulse voltage for all forms des clupment of a discharge during the inception voltage of impulw corona. The of electrodes used, Table I shows the mean breakdown of rod-rod gaps wasinvestigated- corona current, iu =i-fe, diminishes as an vidues for the difTerent electrodes. Voltage and current were measured simul. exponential function. It remains zero for taneously by a cathode. ray oscillograph at u..a voltages at least 20To below the 100To y,y, l the high-voltage electrode. The high. volt
- breakdown voltage Uo of the gap.
-l age electrode was chosen as measuring point so that a single rod, without an op- Electrode ut Nec4rrVE btPL't.sE YO1/rAcz (Fic. .n4) i K r) j posite electrode, could be investigated.
~
Various forms of rods were used; see Fig 22. With u.., voltages 5 0 8 Up the oscillo-The optical view of the discharge figures grams are nearly the same as Fig. 23 with 1.. 2
. + 110. - -T were taken, as are Lichtenberg figures, by inception of impulse corona at fu and .f g ,
photographic paper which was held axully diminishment of the current iu + ic (ie 4.: l + 148. .-95 between the electrrdes. capacitive current). Sometimes the current - The vertically arranged rod-rod gap with impulse of the impulse corona' is totally a distance of 350 mm was investigated abse n t. With voltages 0 8 Uo<u < Uo, All forms of electrodes used clearly show mainly. The distance was always the sam 9 as show n in Fig. 20 a new event w hich the growth of in with inueasing n u, and the peak value of the impulse voltage was is not seen in the cur re nt meillogram therefore with increasing eteepness of the vaned. For all medlegrams the impulse of positive electrales appears. Af ter the impulse vult. ige. With peak values up to voltage of 0 3/40 was used. The itOTo capacitive charging cutrent belonging to u . -300 kv, in is directly proportional to breakdown voltages for the distance of 350 the beginning of the impulse voltage um,; see Table 11. mm were approximately +305 kv and has disappeared, sometimes, during the
-315 kv. return of the impulse voltage at fu, a T& H ve.ry steep and high current ,mpulse, i s,u.
Posittys htetus VorrAcs a ppears. This occasionally is followed by Fig. 23 shows a characteristic oscillo- another, smaller impulse. The impulses ada.as (Amps.100 Kv) gra m. At f i the first part of the impulse disappear as an exponential function. In Electrode + Impuis. - Impu!se generator fires, and at is the impulse voltage every case when the discharge develops to is applied to the gap. During the rise of the breakdown, the current impulse, io, and 1.. .2 2. . .1 35 2., .t.7 . .1.5 3- 135- 1 25
~
i 1 Fie 22. Various fort.s of rod .Iec-F~ l 10 4 10 8 10 4 ' 10 f j - ,
-- - trodes
, Figs. 25( A) and 25(B) shows osciLgrams of voltage and current without breakduwn of the gap, and Figs. 26( A) and (B) with
' " i i breakdown. The total occurrence includ- '
o S l - Y < ing breakdown is shown in Figs. 27(A) and - T I 27(B). In Figs. 28 and 29 the corona gp streamers bridge the whole gap without I_ breakdown. When the breakdown starts I ,N the channels begin again at the electrodes l i B.6 p - k j i .,, j These pictures cannot be reprinted due to , , j 08 'd 2 4 L _ . blackened photographic paper. la the orig-0.2 r U h0.6Af rM j inal photos the formation of the channels is clearly seen. ti n Fig. 23 (left). U Charactr,n.s tic es. T r_ __ o ci!!ogram with F i umox I Uk Uma x ;,p,g,, po,;g;,, l i I Ukl
- t r ,-
O
-et i n I l ".
l -n ik1 1 . ic men f- ' ikt I $4
- t F e,, 24 (risht). _
i , + 1
't 8 3 I Characteristic os- * ' ' ' '
i t3 tzOtu ts2 to ,;iro,,,, ,;,h t3 t 2= 0 ,t u t i,2 to l k-- tit tt2 -H impulse nes.hv. . tw " tt2-" G18 Wagner, Hileman-Mechanism of Breakd<wn of Labora:ory Gaps O cTon n 1Utit j i
. _ _ 9MML22_ :
DJ~ U 'u g ag Je- e MJ. A tru r,..._.--,....,.._.,.m . 3 7.
- - .y -. , .- - - - e ~
r.~ y w, e ...;. - , ; x - . e ) . ,, *. ..7 ,c .. r_ . -;.J >
'm e .."-
s, n r. .'4 . : ,n Vw ps c i (.*
< w - ; .{
l i 2*I i u.Ludm L . ( - d*
- ;.- : . l, ..r ., z 'd ll. ~. ., Ns .:.
0 0. 5 1. 0 1. 5 p s
, .. j ..,.. f. s. }
* / .' .Q . m .'s ;
. .5 >
y'. t
?
' , ? ' , -' .L '
, ',, ' ' ", ' , ' b t,bg . ..D (A) , 0.j , /l7,.', *' ' (
- m , ,s 3 gy, ? ; r l .1.Iy;q gc,
,. g . ;< - .n . ~1 c~ '
.c t :
'1 3 1-
;'. ~...\'.j t, ,
r' < ., j' '
, ~\
..s tN-1
^
l,f l '.. - . \ l't;)'j'ft h! , , f: f r s? ) : /,f iQ k J. ,f s.; Q, { . g , '& , k Y' . ',$!!!*'/.':,,'f)";glyyf ll Q., \ ,)' lI .Y $ '0 l (B) * { "! ./} A;.!
- s,:N'.#)thj\'
g g f j \ [J., gT, ' ,,t . i
. s glll.ffSi',{/(\Y'lf,:
['l/f(s -
,f
; S ,
q O l g/ f, A Fig 25. Osci!!ogrami v' t breakda.n { ! ~ ,' [j ., { i i ' L)( b, iif ts [.
/;l ,N / .',
i
' (' ji,r /J ). .!\ '; /'./ f';i n. (pQ' A -Impulse pcsiti,e h/ ' [' hI f ( \
.. jy '} f g ! } . ()$ ' '\ [ f f,!
B- Impulse negotia, time sco!c as in (A) * 'j 1 i 4 , ,,' s , (. j.(h,i.('y(3 n, , ,. K..I. t' .\ Q c. t- i f. ' q s .h'W g ^; .r h,1\t. ',; n l; h: \ k,r i;k\ ' j ..,, lg , 8 ;
\ <\ , y{;
{- .) ,
- / i ! .) ; e y ~
's . ,
s
/< ./ i 3 '
j ( it . .i ,
. 1 ,
f
&M 4 l' E (' ; ./ -T .{
I ,s . 8 (A) ' >
\, , ,.
, 3
- i. . ; ,i e g ,
I
. \ \ ,!, ,
I
; I
/
)
) ; 3, I \ If . d 1
) 9..
Y
,/ ,
ni 4 , M ' .i ! (B) l, '
,g. '
n- ..-
^
. e Fig. 26. Oscillograms with breakdo.n, time t ( ,.
l l scale as in Fig. 25( A) f -@* .. . " ':
, l , ; ;
A-
- Y., ,
1 , A -Impu!se posmse I + . . - C '
?,tr ". ' ' ,
B - Imodsc 9ega'm _.
,' . ' 'gj* 4 g ,._
{ 4 g
- t.
- t:%.,y.uw; .. : ?:. ' n g r
' V.,.. ,
'. -: ; }
..). ,T:^~i .. '?J. . '. l -? f .,
r '. .3Mi;-.X - f :-
- s. ' m; ; - 1 ..
L, , > . 'Q+: :: -_w f.s ',.f g L:.. m . . - .&. . w4 ./. v _ c._ .._.4 .' , w .2.s _. . (A) i I (A) i l t . ,. 0 $ to ys
/ _ L_ h l 6
o (B) (B) 6
~ ~ g Fig. 28. (A) Positive and negative streamos, Fig. 29. (A) Poutive and negative streamers, .
rod. rod gap, 350 mm, impuhe poetive, rod. red gap, 350 mm, impulse n e g a tiv e, f
'4._ u,,.. = + 2 60 k v. (B) Oscillogram for Fig. u m.. - - 2 'i 5 k v. (B) Osci!!ogram for Fig. i 28(A), time scale as in Fig. 25( A) 29(A), time scale as in Fig. 25(A) l O S 10 p s The sength of the strearners e incasured charge are oven :n a publication about the. !
(g) with one single rod as electrode Here the su bje c t . t length of the corona streamers was propor. The unw rnetho 1 nas used with the rad. ' Fig. 2 7. Oscillograms with breakdo n tional to the peak value of impulse voitage ro.1 gap; see Fig 31. The curana stre m.en (Fig 30 Sorne refketions ur..n the pos. of the positive alectrute bridge a great part i A-lmpuisc pcsitive sible 6 eld strength at the end 4 the enrona of the gap he fore ang'e stre uncrs reach the j 9-Impulse ne ssti<c streamers and the distribution of electric negatn e electrode. l Oc1<ruv n 1Dt1t Wa cw e..'lileman .\lecha nt'sni of Brcak.!xxet of L2 hora.'ory Gaps 610 ,;
8
/t , Fig. 30 (loh). 4 I
Length of posi, lM - 0,}
, u* 30 - - - - - - - -
~~"/7
[ tive end negative [6 corona stre am ers 3 ~ -- ~
, - ^ - " 7 h- ~ - 6 6
3
- z~
y e measured with , , ~~"' O% l vi 51 ' one single rod xw30 -. ---- - 4-- -- 10 $ w 20 # as electrode Q r $ g i w 25 2,
-- 2-o
- 15 %
z , us 20 w o " l 20 - -- - - - - -
$ 10 - -+ ; M --- Fig, 31 (right). o t Length of posi- 3 15 -. .. q. - -
25 5 o _.____-.I l tive and negative o ' t w
' ' I Z
;r. I .5-I j ,,u. gr -y corona streamers u.10 - - - ---
30 u. j - C in th e rodrod O t,00 2 g 0 100 200 300 gap, d; stance 35 5 - - - - - - - n~3 PEAK VOLT AGE IN kV cm 3 o o z l z w, 0 100 200 300 400 9 The time, tu w hich the corona strestners PE AK VOLTAGE IN kV need to bridge the gap and the time, fu, _l cf the development of the channel are taken from the oscillograms, depending on the shown in reference 13, and current oscillo. higher for the spark condition. For the peak value u m of impulse voltage. Figs. grams of Fig. 3. It seems to me that this spark condition also, the tail was sHghtly 30( A) and 32(B) belong to electrode 1. mechanism, as indicated by photographs,is longer and streamer current started directly The main part of the delay of breakdown quite different from the photographs of the at the end of the current pip tail, mcreasing is given by the growth of the channels. Park and Cones' gaps. The photographs s ery rapidly at first and then more slowly. With increasing impulse voltage the time, of the latter paper, obtained with chopped in a concave manner, to aparkover current tu. decrews very rapidly. Neurt heless, n aves appli-d to a sphere-to-plane ga p, with sparkover being con.pletid in 15 see it is at least twice the time, fu. indicate ntnnerous streamers but do not or a total of S psec from current tero, The gap rod plate as well as the rod-rod show the glow discharge shown in the 000 Since it was indicated that the streamers gap was investigated. With positive im- inch gap, appeared to meet in mid-gap, the ground pulvs the dictance was 20 mm, with nega- 1.'nfortunately, one cannot he certain of a strea:ner vehicity would be approximately tive impulvs M nun. Both show the same correct im e rpret a tion because only si dl 1.7 X 10' em 4ec u hich is similar to that c haracteri* tics in principle. The irhence photogt iphs are avaitable. Ne s ert heless, given by the authors' equadon C.
- cf a resistance at the high voltage electrale my intstpri tation, based on photos and in the ime rmediatc condition w here the j was also investigated. Details will be mciDograms, is as idluws: The Oto inch gion brniged the gap md a streamer w as f..und in a paper soon to be published in gap voltage was just below the critical (10% seen, a second current pip dewb ped about sparkuver) resulting in nine impuhes uhere 8 pec af ter the first pip was reduced to zero, the Fleckhetechniuhe Zrhhnfl. in t he case where the glow reached part a ne:ik, ultraviolet glow Spanned about half of the gap, and current trace B of the w ay, about three quarters across the gap, J. H. IIagenguth (General Electric Com. authors' Fig. 3. In nine other caes the there was no second current pip, or if there pany, Pittsfield, Mass ): This paper is an glow bndged the gap and a streamer de- uere, it w as le=s than 1 ampere, the limiting escelkut summary of t arious laboratory veluped from the grounded rod, current sensitivity of the current measuring eircuit.
work on the sparkover of nonumform field trace C; and in two cases parkover was A very short, small gl ov at the grounded gaps. However, I was shghtly confused completed and resulted in current trace D, rod appeared. by the interchangeable use of the terms In these latter cases there w as a well defined The very heavy glowing ball from quartz space charge, corona discharge, corona split in the spark in the middle of the gap, lens photos around the exated negative efec. streamers, channels, and discharge thannels. indicating where the final streamers emanat. trode appeared as very thin streamers of the One cannot always be certain of interpret. ing from both electrodes met. In all three ume lengths when photographed with a ing these terms currectly. conditions there was an initial current pip. glass lens snmlar to the Park -Cones' i The authors discuss the sparkm er process This is ahnost equal in the two cases where phot os.
, of a .'00-inch rod rod gap at 0.too kv, as the spark was incompleted, and about 2u% Thus it appears that this glow may be 35
! 3.5 l , _ - _ _ . _ _ _ _.
3 . 3 4 l 1 !
! d 25 - -
vw 25 - - - - I m I o o m l l - - . - - --
. 2 -
7 2 - b 9 {
)
i
- 1. 5 - - -
(-- z 1.5
- - - + - - -
I s tn l ---{ --- I
$1 __ _b Fig 32. Time of
" 3 g I t2 7 1
! l 1
strea rn er and 0.5-- 15 ------ -~ ~ cEannel for 35
]- -l ~ +
! 0 O A P oIa ri t y 200 300 30 200 300 20 d'
PEAK Vol.T AGE IN kV PE AK VOLT A 3E IN kV 9 P a'. ' f r34 (A) (E) tis e G20 IVqm r, Hiirman .m danism of Br.ddm of Ldora:ory Gaps Ocm n M1 90004124
G- An ksom .t f FE as the electrons produced by the iodzation k simit ir to the pilot leader po,tulate ! by R oW s. H E. Le 6"l . J Si hviarid From the li nited data there ar- at fe ist t w o pos,iNe proctues for this
$['Q Q( ['y * " d "d proce nes at the tip of the streamers are draw n inward toward the rod and the urn!< - ( dt. t , ev e nliti"rw
((."",I7'['Fm c ']
,, Q 'y ' "'[ Ra .
photigr4 hic recorJ sh me i m ty convit ite a true n- aute of the b .undry >f the ten T n. , .w.. un. r> c m ,' a L 1. h . pi .t k hr tar's at the esa r_, ,yg s y, im , 7M me i n s t h '.t t e rJ n How .tr neg ui.c r. .d tow 2r.! t he c w N d p. ,i t n e 1 a with ne g it n e pot e:.nal. the elm r' m rod, de ni~pimg a ver y weak, high c.e,. c, n g,ecy,t roep ny, u t,ot sm i %u n, t s .i.
,,, y rif,att or N s r. h , formed in the ioniz nion pr ness h2ving hii;h ans plav .a (order of 27xle ohn#, x, . w,, o y y , m ,, n. % tm rnobiHties can .elvance outw an! bepnd the bru!gm the op, and a :,trcamer starts from 4 s u c n u s u. or rue sasar oi,c es sii ;. . J. bou nca ry indicated by the phot'sg aphic the grounded rod. This streamer bndg% w. no.e n, fw s m n mi a oct m i, reewd. %'e m ww w hat reluctant to en ,
ab gt c of the gip ind y et no 3tream r pp A" e- idon on the gw ous c'ectrvnic is vmke from the nepth e e!ectrode. 4, , y , m m wi thMm m d this 2 T he pdot leader starts at the escited C. F. Wagner and A. R. Ili:eman : T h.- Pd "t "I *" " ' I" ' h" " "'" "I C"" I negatis e rod toward the grounded positive ture rather than fact. , data presented by Profew ir Daatz and rod and is met by a sunil tr leader from the m FMer ne vers illunanatmg and shi There is a dhence M 5,nion ewerning ; grounded rod- Af ter contact, a streamer the nugmtude of current involved m the , I ennuderably to the knowlede of the 5"P' "I '.he 4,a ning stroke The d ita starts froni the grounded rod, as in I subject. The authors look forward with ilN e ver, it is p. oiMe that the le ider from presented m TaNe 11 of the divuu m m ty I gie it interest to the Ek%tn hin h &it. shed some light ,n this point. It is stated the r undrd ru! is a streamer, although g,yg ,g j g % gg g, g -nan rod rod g ip for neg stis e e
,, s B and C 1. not wem to mdic ite wNd'm m &td Tb Mmq mm th"' I"r i 3
"" b E' N P"P' '
- ments ue directed to th.ctr s erv c! ear but 3 P N.
- to L up to J. 00 kv and les than 2 im.
nemsardy limited prhentaten. Nnce the The se interesting conclusions state that in this 200 inch rul g sp a leader deselops prunary purpose of t! e re vie w by the a M d MM W m 4 across the entire gap withuut a streamer authors of the available i laboratory data l j advancing t<nurd it from the ground elec. w as t provide a backgr und upon a hich to g base 2 theory of the .ightnin g discharga, trode, es en though the e!ertrode 5 n a non- me matedd pn eno"I W Pr .fes..r Da itz , aualle referred to as t he .ye u ru f .r ~- i tl araund a" The riur-tv n then
,~ ' 'o ,# 'h 4 * "'P' d return stri e '
a ha t L a si : d ir pre er occurs m the 2"d D ' to t hu tin d of *"4. ' i Jghtnw g ,trile The Antheirs regict that Mr. Hagenguth , W ith reg ird to average gap gr adie n t s. w u confuel bv the interchang of terms. l 8mphon Tnd an !)sa I the iteps nf the g the authors arrne at a 64ure between b'M arst con einent > t a Ughtumg strose is the s and lu'iN!O volt s 'c:t;' w tuch is correct for used dir'erent terna (< >r the. >tme phenon,r , elect ric gnde "t within the sp ire chare ~ l iropu6 appootion uni the ty pes of gaps enon in order to ditie rc u tat e be t w ee n , stu M! In a re"ent n in r i unu .u d, low
- 5"" 'u ndin g the condn tin g core of the two qtute ddierent phenomon t.
le/* - Or m-p re I m et he miy. the I dry .o, i 3 h n . r s t r e q % 'a r e 3nown in I ig Mr- H g.m gu t h pr e u 4 an inten dr >- radient along the streamers that , entute
- 21. At m0 inchn the n era ":s gradici.t ., d2 scu d .n conerning tha occurre s . of the corona duc, a arge Ftg 00 prmido in-was IMO voit Cc n' Them !!a3 hovers w ere streamers that appear to bo.dge the entire obt aine d w ith slow -frunt switching surge- formation on this point- In the pre'imin try We are of the opinion that this pp ty pe w ases and also with fast front long tail c pv of the diwunion it was stated that . . ,
phenomenon}}