Experimental. E-Gun. E-Gun Modulator Arrangement AI VI MONITORS TRIODE ELECTRON BEA~ CATHODE TRIGGER

Transcription

1 TEMPORAL WAVESHAPNG OF A TRODE COLD CATHODE ELECTRON BEAM GUN Gary Loda Beta Development Corporation, 557 Sierra Lane Dblin, CA 945 Sol Schneider* Little Silver, NJ 7739 William F. Otto and George J. Deenberg US Army Missile Command, ATTN: DRSM-RHS Redstone Arsenal, AL ntrodction t as first shon by Loda1 that the effect of plasma closre on the impedance of cold cathode diode electron beam gns2 cold be greatly redced by adding a self-biased grid to the gn. Large triode cold cathod electron beam gns are being sed by Los Alamos National Lab~ratory in the ANTARES system.3 A characteristic of the triode cold cathode e-gn is that a large initial crrent overshoot occrs as the self-biased grid charges to the operating voltage and the acceleration voltage has a poor rise time hen a plse forming netork is sed for excitation. Since foil heating is inversely related to e-gn voltage, the voltage rise and fall times shold be minimied to redce the heating. Leland and Kirchner recognied the need for improving the initial triode e-gn transient and sggested sing axiliary circitry for rapid readjstment of the grid potential after cathode ignition, sing a "spark" cathode or redcing the grid-anode capacitance.4 n this paper the characteristics of a 15 em by 2 em beam area "spark" cathode, triode electron beam gn are presented and the modeling and experiments performed to improve its voltage rise time are described. E-Gn Experimental The cross section of the electron beam gn investigated in this ork is shon in Figre 1. A perforated stainless steel cathode strctre, configred in a Pierce profile, as sed to facilitate investigation of different ide area emitters. n this ork a "spark" cathode as sed ith the emitters being 48 separate resistively ballasted srface flashover cathode discs. The discs ere arranged in a single ro ith a 4.1 em disc separation. The concentric grid as made of solid sixteen gage stainless steel and fitted ith an 8% open area molybdenm screen. A field shaping ring as affixed to the grid to improve e-beam niformity. The separation beteen the cathode and grid planes as ell as beteen the grid and anode planes as 15 em. A copper hibatchi strctre ith 57% open area, sealed ith a 1-mil Ti foil, as sed as anode. Modlator The e 1 ectron beam gn is driven by the 1 i ne type modlar shon in Figre 2. Three parallel netorks ere sed to facilitate PFN impedance changes for different modes of e-gn operation. Ten stages ere sed per netork ith a capacitance vale of. ~Fd/stage. The stage indctances ere individally adjstable from to 15 ~H to provide for plse shape adjstment. *Army Research Office/Battelle Colmbs Laboratories Scientific Services Agreement 48 Figre 1. Triode Cross Section A to electrode, ferrite decopl ed, spark gap as sed to sitch the PFN. An inverse diode and an end of 1 ine clipper are not sed since a spark gap is a bidirectional crrent condctor. Ten parallel 9.3 m long 5 S"l cables ere sed to connect the PFN otpt to an oi 1 insla ted 17: 1 step p plse transformer. The transformer secondary is conected beteen the electron beam cathode and the gronded anode. A trigger transformer is connected beteen the cathode and the 48 resistively isolated cathode emission sites. The trigger transformer is 1 ocated in the oil tank ith the plse transformer and is isola ted from grond for the fll 3 kv maximm cathode operating vo 1 tage. Pl sed vo 1 tages in the range 5 to 25 kv ith. 5 ~sec pl sei dth are sed for cathode i ni ti ati on. Less than 1 mj of energy is reqired for each cathode srface flashover site. The modlator contains a charging diode and indctor to permit operation ith resonant charging p to a 5 H rate. A large 2 kv poer spply as na vail able and the experiments reported here ere performed at a repetition rate of seven seconds beteen plses. Figre 2. A V MONTORS E-Gn Modlator Arrangement TRODE ELECTRON BEA~ CATHODE TRGGER

2 Report Docmentation Page Form Approved OMB No Pblic reporting brden for the collection of information is estimated to average 1 hor per response, inclding the time for revieing instrctions, searching existing data sorces, gathering and maintaining the data needed, and completing and revieing the collection of information. Send comments regarding this brden estimate or any other aspect of this collection of information, inclding sggestions for redcing this brden, to Washington Headqarters Services, Directorate for nformation Operations and Reports, 1215 Jefferson Davis Highay, Site 124, Arlington VA Respondents shold be aare that notithstanding any other provision of la, no person shall be sbject to a penalty for failing to comply ith a collection of information if it does not display a crrently valid OMB control nmber. 1. REPORT DATE JUN REPORT TYPE N/A 3. DATES COVERED - 4. TTLE AND SUBTTLE Temporal Waveshaping Of A Triode Cold Cathode Electron Beam Gn 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNT NUMBER 7. PERFORMNG ORGANZATON NAME(S) AND ADDRESS(ES) Beta Development Corporation, 557 Sierra Lane Dblin, CA PERFORMNG ORGANZATON REPORT NUMBER 9. SPONSORNG/MONTORNG AGENCY NAME(S) AND ADDRESS(ES) 1. SPONSOR/MONTOR S ACRONYM(S) 12. DSTRBUTON/AVALABLTY STATEMENT Approved for pblic release, distribtion nlimited 11. SPONSOR/MONTOR S REPORT NUMBER(S) 13. SUPPLEMENTARY NOTES See also ADM EEE Plsed Poer Conference, Digest of Technical Papers , and Abstracts of the 213 EEE nternational Conference on Plasma Science. Held in San Francisco, CA on 1-21 Jne 213. U.S. Government or Federal Prpose Rights License. 14. ABSTRACT 15. SUBJECT TERMS 1. SECURTY CLASSFCATON OF: 17. LMTATON OF ABSTRACT SAR a. REPORT nclassified b. ABSTRACT nclassified c. THS PAGE nclassified 18. NUMBER OF PAGES 4 19a. NAME OF RESPONSBLE PERSON Standard Form 298 (Rev. 8-98) Prescribed by ANS Std Z39-18

3 Diagnostics The electron beam gn cathode to anode voltage as measred ith a Stangenes Model CVD-2 capacitive voltage divider and the total cathode crrent as measred ith a Pearson Model 11 plse crrent transformer. The monitor locations are shon in Figre 2. The signals ere roted on coaxial cables to a screen room and recorded on a Nicolet Explorer digital oscilloscope. The data as stored on floppy discs for ftre analysis and compter processing. v L5 V Cl CATHODE t v GRD ~ V3 ANODE Figre 3. Triode E-Gn Model E-Gn Model The lmped element e-beam gn model sed is shon in Figre 3. The crrent sorce A4 represents the cathode emission and is described mathematically by the Child's La expression: A4 = X 1- A V23/2;d2(1-T/T4)2 (1) here: A is the e-beam area V2 is the grid to cathode voltage d is the grid-cathode separation T is the time variable T4 is the closre time The closre time T4 is eqal to the ratio d/v here V is the cathode plasma closre velocity. The crrent X A4 in Figre 3 represents the crrent transmitted throgh the grid, here X is the fraction of the cathode crrent transmitted. X is abot 8%. VO and RO represent the open circit PFN voltage and PFN impedance, respectively, reflected to the transformer secondary. The series indctance LS as determined by measring the shorted transformer secondary inding crrent aveform and matching the risetime ith a LS/RO time constant. The capacitance vales of C2 and C3 ere determined from bridge measrements of the grid-anode capacitance ith the cathode to anode circit open and shorted and conversely. These vales ere C2 =.75 nf and C3 =. nf hich are close to the calclated vales of C2 =.5 nf and C3 =.5 nf. The calclations did not consider the bshing nor external monting capacitance. C1 =.3 nf and is composed of the sm of the e-gn inpt capacitance and the.175 nf plse transformer stray capacitance. The grid resistance, R3, as a liqid resistor hose va 1 e as adjsted by changing its copper slfate concentration. R3 vales ere determined by applying a 115 V, H sorce and measring the reslting crrent. The closre time T4 as eqated to the measred time from e-gn voltage onset to sdden voltage drop and crrent rise. This is illstrated in Figre 4. This sdden voltage drop did not occr on every plse bt hen it occrred, the time to voltage drop as consistently beteen 8 and 9 ~sec. The vale T4 = 8 ~sec is sed hen the calclations are compared ith experiment. Dring initial e-gn experiments no closre as noted for plseidths exceeding 1 ~sec KV,2A Figre 4. E-Gn Voltage and Crrent With Closre The case of the closre illstrated in Figre 4 is not nderstood and is being actively investigated. A program as ritten for the Helett Packard Model 85 Personal Compter to solve the eqations describing the circit shon in Figre 3. The reslts obtai ned for three sample cases agreed ith the otpts of SUPER-SCEPTRE and TRACAP circit analysis programs.s The temporal behavior of the voltage V1 and crrent Al as calclated for different e-gn operating conditions and compared to the corresponding experimentally measred vales. Reslts Sensitivity Analysis The circit parameters shon in Figre 3 ere varied to determine their i nfl ence on e-gn impedance and transient behavior. The components connected beteen the grid and anode had the greatest inflence on the e-gn performance and the external components VO, RO, L5, and C1 had the least inflence. For the cases investigated the e-gn impedance as approximately constant for VO vales beteen 2 and kv. Bel o 2 kv the impedance rose gradally to 2% above its steady state vale at VO = 5 kv. Also, the e-gn impedance as i nsensi ti ve to the beam area A of Eqation 1. The e-gn impedance only decreased by abot 3% hen the beam area as redced to 1% of the fll apertre vale. RO, L5, and Cl did not stron~y effect the steady state impedance bt did effect the transient behavior. The e-gn voltage rise time increased ith increasing vales of L5 and Cl. Hence, these circit elements shold be minimied. A seqence of figres ill be sed to illstrate the e-gn performance dependence on the circit components shon in Figre 3. Unless otherise stated the folloing standard conditions ill be sed in the figres: VO = 4 kv, RO = 843 ohms, LS = 1.2 mh, Cl =.3 nf, C2 =.75 nf, C3 =. nf, X =.8, T4 = 8 ~sec, A = 3, cm2, d = 15 em, and R3 = 59 rl. The e-gn impedance is primarily determined by the vale of the gri d-andode resistance R3 and the grid transmission fraction X. The e-gn impedance monotonically increases ith increasing R3 vales, as illstrated in Figre 5, and e-gn closre redces the impedance from the no closre vale. n Figre 5 the no closre case as calclated by 1 etti ng T4 = oo and the impedance ith closre as obtained by calclating the ratio of Vl to A1 at. 5 ~sec ith T4 = 8 ~sec. The impedance vale at.5 sec is sed to facilitate comparison ith experiments. The difference beteen the to crves in Figre 5 is exaggerated at large R3 vales de to rise time effects. The steady state impedance ith closre is not achieved at.5 sec at 1 arge R3 vales. The e-gn impedance decreases ith increasing grid transmission. This is illstrated in Figre for the for vales of grid resistance R3 investigated in this ork.

4 4 :::: ::S::3 2 ( 1 :::: NO CLOSURE SEC CLOSURE > GRD-ANODE RESSTANCE R3 N KOHMS Figre 5. :::: ~5 ::s:: >-4 ~3 D ~2 1. E-Gn mpedance vs Grid Resistance R3. R3=11 R3= GRD TRANSMSSON FRACTON X Figre. E-Gn mpedance vs Grid Transmission Fraction X. The e-gn voltage rise time behavior is illstrated in Figres 7, 8, and 9. n Figre 7 it is shon that the voltage rise time increases ith i nreasi ng vales of grid-anode capacitance C3 and illstrates the need to minimie this capacitance.4 Unfortnately C3 is difficlt to redce becase the major contribtor to this capacitance is the vale associ a ted ith the concentric grid and anode cylinders. Hence, other means mst be sed to redce the e-gn voltage rise time. The voltage rise time can be improved by adding capacitance beteen the grid and cathode. This is illstrated in Figre 8 here it is shon that the voltage rise time can be improved by a factor of to if 4 nf is added to the residal C2 vale of.75 nf. 5 D ~4 ::: i3 >2 ::s:: [[) ::::;1 D +-~---r--~--~-+--~--+-~---+--~ :::: ~ GRD-ANODE CAPACTANCE C3 N nf Figre 7. Voltage Rise Time vs Grid-Anode Capacitance C ~3 ::: >- U2 :::: f- o +-~---+--~--r-~---r--~--~-+--~ ~ r- ADDED GRD-CATHODE CAPACTANCE N nf Figre 8. "3 :::E '-' ~2 ::: :=J ~1 '-' L') _.J > Figre 9. :::E 1 ~ ::E 2 :=J L') Voltage Rise Time vs Added Grid-Cathode Capacitance Al;C2=.75 nf /'-.1; C2=2. 75 nf 1'.1; C2=4. 75 nf Vl;C2=4.75 nf Vl;C2=2.75 nf Vl;C2=.75 nf TME N MCROSECONDS E-Gn Voltage and Crrent Variation For Different C2 Vales. +--r_,--+--r~--+--r~~~-+--~~~ Figre TME N MCROSECONDS E-Gn mpedance Variation For Different C2 Vales. The temporal behavior of the e-gn voltage and crrent for different C2 vales is shon in Figre 9. The voltage rise time improvement for increasing C2 vales is apparent. t is seen that for C2 = 2.75 nf the voltage is fairly flat and for C2 = nf a vo 1 tage overshoot occrs. A 1 so ill strated in Figre 9 is the decrease of the i ni ti a 1 crrent peak and the higher va 1 es of 1 ate time crrent occrring hen C2 is increased. The

5 temporal variation of e-gn impedance is i 11 strated in Figre 1. t is seen that a moderation of the e-gn impedance variations can be accompli shed ith a proper choice of C2. Very flat impedance profiles ere calclated for X=.7 (i.e., redced transmission from the standard case) and C2 = 2.75 nf. Experimental Comparison The measred and calclated e-gn impedance vales are compared in Table for for different grid resistance vales. The measred vales had a~ % deviation from the mean vales listed in the table. f the extremes of the error bonds are sed reasonable agreement beteen theory and experiment is obtained for a grid transmission vale beteen. 78 and. 79. Hoever, if the impedances in the i ndi vi da 1 ros are matched, increasing va 1 es of grid transmission, toard the geometrical limit of.8, have to be sed as R3 increases. This may be a real effect becase at constant gn crrent the grid to cathode electric field i 11 increase ith increasing R3 vales. The increased electric field cold alter the e 1 ectron trajectories in the cathode-grid region. Table. Comparison of Measred and Calclated Vales of Gn mpedance, Z1, For Different Grid Transmission Fractions, X, and Grid-Anode Resistances, R3. GRD MEASURED CALCULATED Z1 (Sl l R3 (Sl) Z1 (Sl) X =.8 X =. 78 X =. 77 X =.7 1, , ,9 1,12 1,47 1,178 1,24 1,315 11, 1,89 1,7 1,923 2,58 2,199 TDK, ncorporated 1 nf, 5 kv ceramic capacitors ere added beteen the grid and anode to improve voltage rise time. The experimental and calclated temporal voltage behavior is compared in Figre 11 for C2 =.75 nf and C2 = 4.75 nf and the temporal variation of crrent is compared in Figre 12. For C2 =.75 nf. Good agreement is obtained beteen the calclated and experimental behavior. (f) 1- _j 3 > C2=4.75 nf ~ ~2.-- > ;s 1 _j C2=.75 nf > TME N MCROSECONDS Figre 11. Comparison of Calclated and Experimental Voltage Variation. 489 (f) CL :::E: 3 > ~ 1 C2=.75 nf TME N MCROSECONDS Figre 12. Comparison of Calclated and Experimental Crrent Variation. Figres 11 and 12 ere generated by plotting the otpt of the Nicolet Explorer Digital Osci 11 oscope directly onto ca 1 cl a ted crves. The grid transmission fraction X as taken to be. 78 based on the data shon in Table. The open circit voltage VO as adjsted in the calclations to match the experimental data. The folloing vales ere sed: VO = 37 kv for C2 =.75 nf and VO = 3 kv for C2 = 4.75 nf. The charging voltage as not measred in the experiments. t as adjsted dring each experimental condition investigated to prodce e-gn voltages of abot 2 kv. The comparison of calclated and experimental crrents for C2 = 4.75 nf is not shon de to page restrictions. The experimental data agreed ith the shape of the calclated crve in that a central dip and a positive end of plse slope as measred. (See, Figre 9 for calclated shape.) Hoever, the amplitde agreement as not as good as for the C2 =.75 nf case shon in Figre 12. For C2 = 4.75 nf, at J1 sec the calclated vale as abot 13% loer than the experimentally measred crrent. Conclsion The model shon in Figre 3 as fond to adeqately describe the general behavior of a ide area triode electron beam gn. Good agreement beteen theory and experiment as obtai ned of e-gn impedance dependence on grid resistance. Also, the predicted voltage rise time improvement obtai ned by increasing the e-gn grid to cathode capacitance as experimentally verified. Model refinements are needed to describe the observed high freqency voltage and crrent variations. References 1. G. K. Loda, Proc 2nd nt'l Topical Conf on High Poer Electron and on Beam Research and Tech Vol, Lab of Plasma Stdies, Cornell Univ pp , Oct 3-5, G. A. Mesyats et al., in 4th nt. Symp. on Discharges and Electron nslation in Vacm, p 82, Sep 1-4, W. T. Leland, et al., LANL Report No. LA- 718, Jl W. T. Leland and M. Kirchner, LANL Report No. LA-UR R. Limpaecher, Sper-Sceptre; J. R. Hall, TRACAP Comparison.