Cathode Beam Behavior in a High Speed Oscillograph Citation Harrison, Arthur Elliot (1940) Cathode Beam Behavior in a High Speed Oscillograph. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/ptan-cr41. https://resolver.caltech.edu/CaltechTHESIS:12172024-211553369 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Electrical Engineering) Degree Grantor: California Institute of Technology Division: Engineering and Applied Science Major Option: Electrical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Sorensen, Royal Wasson Thesis Committee: Unknown, Unknown Defense Date: 1 January 1940 Record Number: CaltechTHESIS:12172024-211553369 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:12172024-211553369 DOI: 10.7907/ptan-cr41 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16936 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 18 Dec 2024 19:37 Last Modified: 18 Dec 2024 19:43 Thesis Files PDF - Final Version See Usage Policy. 29MB Repository Staff Only: item control page

Cathode Beam Behavior in a High Speed Oscillograpb Thesis by Arthur E. Harrison In Partial Fulfillment ot the Requirements fop The Degree ot Doctor of Philosophy California Institute of Technology Pasadena, California 1940 TABLE OF CONTENTS INTRODUCTION ............................ ............. . ,. Page l EQUIPMEm • •• • •• • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 3 PHOTOGRAPHY •••••••••••••e•••••••••••••••••••••••••••••• 13 DEFLECTION DISTORTION•••••••••••••••••••••••••••••••••• 16 BEAM DISTORTION•••••••••••••••••••••••••••••••••••••••• 10 Elliptical Diatonion •••••••••••••••••••·••••••• 10 Tl!'t,uud.ent G1.1 Focusing •••.•.••..•.•.•.••••••••• • 26 Bead Distortion•·•·•·•••·•·•••••·•·••••••••••••• 36 AN·OTifiALIES • ••••••• • • • ••• • • • ••••••• • ••••• • • • • • • • • • • • • • • • • 39 ELECTRON OPTICS•••••••••••••••••••••••••••••••••••••••• 44 Opties •• •••••••••••••••••••·••••••·••••••••••••• 4t Beam coneentn.ti.011 •••••••••••••••••••••••••••••• 47 BEAM STUDIES••••••••••••••••••••••••••••••••••••••••••• 62 Etteo-t ot Water Va,or ••·••••·•••··•••·•••·•••••• 62 18 Cyclic Variations or Ettio1eno7 •·········•·••••• 5'7 Cathode Sputt•ring •...•.... . ...................• • 58 'WRITING SPEED • •• • •••••••••••••••••••••••••• • • • •• , • , •• • • 69 RECOMMENDATIONS ••••••••••jJe 9·• ••••••.,•••••••••••••• ••••• 63 BIBLIOGRAPHY •.•••.•••••••• o • • • • • • ,,, • o • •••••••••••••• • • • • 66 ACKNOVJLEDGMENT The assistance of'f'ex-ed by Dr. R. w. SoNnsen and all membe:r-s ot the Eleotr1ea.l Engineering faculty of' the California Institute ot Technology is very grat•fully acknowledged. Dr. Fred Lindva.11 should be thanked in particular for constant encourag~nt and frequent advice . .A suggestion by Dr. R. B. Brode at the Univenity of California led to the explanation of trueient gas focusing. Dr. Howard Grieet authoriHd the use of' certain illustntions ~om his thesis to describe the equipment used, and Bernard Oliver and v. C. Rideout often assisted in obtaining data. S'UMN..ARY High speed cathode-ray oseillographs utilizing high vaouum tubes and external photography are now available, but oerta.in eleotrloal N• searoh problems l'"equi!"e writing speeds in excess or those possible with suoh equipnen-t .. A moderately low voltage, oold cathode type, high• speed oseillogra.ph for internal photography is part or the Institute High Voltage La.bon.tory. Its capabilities beyone the l"a.nge possible with the present high nouum tubes have been investigated. The :taotor• affecting its operation were studied, and tha results of this study have suggeeted improvements in the equipnent, as well a.s furnishing data of considerable pre.otioal value and theoretical interest. Dietonion a.ftect:i.ng the magnitude and direction of defleetion of the beam has been •liminated and distortions affec,ting the d.&e and shape or the beam spot have been reduced. Transient gas toousing in the tube has not been eliminated, but the tactors causing transient gas focusing have been studied, and a tentative explanation of the phenomenon is suggested. The operation of the osoillogn.ph, as an electron mioJtO• scope 11 described. The principles of eleotr-on optics ar-e used to•• plain the observed results when an axial magnetic field 1s used to tn. orease the ratio of the beam current to the total discharge cu:r-J>ent. The effects of gas and Taper pressure, ea.thode surface conditions and cathode age on the operation of the oaoillograph are described. The writing speed of the osaillograph 1V!ls ertended e01llewhat and was measured as 2 x 109 om. per aeoond.. Published papers made possible by the im- proved operation of the osoillograph are included as an Appen.db:. INTRODUCTION Dependable high speed osoillogra:ph equipnent is an essential part of any electrical testing laboPatory designed for high voltage measurements. several high..speed osoillog:n.pha of' the oold cathode type were oonst:ruotei and operated in the Institute High Voltage Laboratory prior to the availability of oommer-cial high-n.ouum oaoillogPa:ph tubae tor high apeed recording with an external oani.era. Theae oold. cathode tubes ha.ve been described in dootor•s theses l>y Grieat1 , flawley-2 and Plea1ant1I. Although high vaouum tubes with s~ftioieat Wl"iting speed torlightm.ng sUl'ge studies a.re now available, then i• still a range ot open.t1on whioh is satisfied best by a. ttold oathod• type oscillegraph with internal film Noording. The tube designed by Dr. Howard. Griest was chosen tor further development in 1936 beoauae it was Pelatively simple to operate and required considerably leH aooelerating potential than other tubes used previou•ly. One of' the principal uses of a high •peed osoillograph 1s the i-eoording of tl'anaianta tollcwing the application or &rtifioia.l lightning surges to eleetrioa.l equipnent. A surge genei'ator whioh oe.n be tripped meohanioally is used in the Institute High Voltage Laboratory for im• Suell a system requires an electrical connection between pulse testing. the surge geneN.tor and. the osoillograph for coordinating the tvn, set• of equipnent. The surge tripping meohmism, oscillograph trip oen• neotion and delay oable cont1eot1ng the voltage divider across the te•t apparatus and the osoillograph dei'leotlon pla.tee must be designed so the entire transient phenomenon may be Ncord•d. Also, the oscillogra.ph should not be disturbed by the electromagnetic field ereated by the surge generato21'. These requirements al"e diftloult to attain completely, l but a satiataotory comprt.1nise was effected so tha.t l"outine teet1ng waa possible. Since impulde nseareh as well as commercial testing wa.s in pioogtees almost continuously, it was necessary to carry out a development pr-ogram without intertering with routine testing. At the same time, oathGde beam behavior was atudied in order to improire the oscil1og1"am recol'dliJ. As a result., the oathode beam studies weJ'e often intimately !"elated to the development of opera.ting teohnique* so that no attempt hae been me.de to separate the two t&.otol"s, a.lthou~ the results may "8.t'y, r~om electron phenomena which are mel"'ely inf'orma.tive, to oha.ngee whioh concern purely meehanioal improvements 1rt the equipment. No radical changes i n the de- sign or the osciillogra.ph were considered. The VJOrk waa limited to an investigation of th& h.otors affeoting the optimum e>peration 0£ the existing equipnent., EQUIPMENT A continuou1 cathode beam. is furrtished by a metal di$<maJ"ge -tNb• using a removable aluminum oathoda and an anode with a 0.016 inch diameter window. Air is admitted by a needle valve leak to allow a cold dis• ohe.rge. The neeclle "f'alve lead ecr-ew is drilled so the leaking gas or air can be controlled or passed through a drying tube u• necessary; and the lead screw is lubrioe.ted with stopcock: grease to prevent an inlet of air at that point. Figure 2. Tha details ot the cathode assembly are given in. A blocking chamber of Rogowski type is used to pre-,,ent fogging of the film while n0 l"ecord ie being made, Additional data.Us of the 1 design of the osoillogra.ph are given in the thesis by Dr. R, H. Grieet • Focusing of the beam is aoocnplished w:l th two oona mounted ia iron. cases with a small air gap to give a oonoentrated field oouial with the cathode beam. One ooil is looated above the s,..veep and detleotion plates, approximately midny bet?Jeen the anode window and the ao:reen, to tom an image of the anode window at the Sol"dn and give the sha~at possible tl"aoe:. The other eoil has fewer turns or la.rge:r wiNJ it la located just above the anode, and is used as a. f'or-e-oonoentration ooi1 to inoreas$ the number of eleotrons in the oathode beam. Current to~ both ooils is obtained f:tom a. ai.x...volt storage battery oont?"olled by potentiometers. Deflection normal to the sweep direotion is causi,d by a pai~ ot deflection plate• approximately 40 cm. above the tluoteao~nt 1or••n• A selector switch mounted near the defleotion plates eonneots these plates to different taps on the te:rminat1.ng impeda.noe of' the delay oable for impulse ta.ting. An elaotron coupled radio-frequency oeoillator furnishes a timing wave. Deflection sensitivity is measured with ad-a 3 voltage supply oontrolled by a set of tapped resistances and a 1000 ohm rheostat in series with the primary et the po•r transformer of the ~eotifier. A reversing switoh ~or the high voltage supply was built into the ts.p switoh on the oontrol panel but was not oonrteoted in-to the cireuit. A set of cons connected through a reversing; awitoh to a. potentiometer' a.cross the six ...volt ston.ge i:>attery furrd.shea e. magnetic bias to adjust the position of the oathode beam in the detleotion direotiOli. Six d.if'f'erent s-weep speeds from less than one miorol$CO!ld to e,evera.l hundred mioroteoonds duration ca..n be eeleoted by changing a tap switch. The sweep impulse is capacitively coupled to the swep plates so that electrostatic bias oan be used to control the position of the beam on th8 so:reen. ·• A "heating circuit" conaisting of sevel"e.l condensers and some nsistance is often used to improve the spark a'\ the trip gap. Such a cirouit was discarded a:f'ter a trlal because u improvement was apparent when fast $weeps weN used,; and th& circuit actually caused periodic reblocldng with the alower· sweepa. Several 400 ohm resistors are requir1!d for damping oscillations in the LC oirouits f'o!'mad by the dii':ferent plates and their leads. 400 ohm resistors a.re not quite high enough for oritioal damping, but they red~~ the ~soillations ad•quately. Very high rrequenoy osoilla• tions are apparent in the detlaction plat$ o1rouit with the fastest sweeps, but they are small in amplitude end not objectionable, although they a.re relatively undamped. Other circuit Yalues for the blocking and s"1eep circuits are given 1n Figure 3. Figure 4 givee the connection.a for all oircuits on the main control panel. No change was ma.de in the oamera and screen meeha.nism orig1na11y oonstructed by Dr. Griest. A description of this mechanism will not be repeated here. but a photograph of the camera is 1noluded as Figure S. 4 FIGURE 5 Table I Terminal Numbers Circuits 1• 2• Earth's Field Compensation 3 • 4 • Focus 14 Fore-conoentration Magnetic Bias Battery Elect:rostatie Bias Filam~nt Eleotrcstatio Bias Plate Smep Filament Sweep Plate 15 16 A. C. Gi"Qund Fo?"e PtUllp 5 • 6 • 7 • 8 9 -10 • 11 12 13 17 Vapor Pumps 18 19 Kenetzoon Plate 20 21 22 Kenetnn Filament A. c. Input Glow Tube A. C. Master Swi toh 23 Sensitivity Bias Plate 24 Not in use • 2 • 4 - 6 • 10 D. c. Positive and common lead. 10 The osoillogra.ph is oonstruoted of. braes and requires a coil sunounding the entire tube to nullify the Effe<Jt of the earth• i! magnetio field. If not compensated, the magnetic tield of the earth causei a derleotion ot approximately one inch and may deflect the disoharg• suf'fioiently to reduce appl"eoia.bly the number- et electr-ons paesillg through the limiting apenure below the anode window. Three stages of' meroury diffusion pUYUpt and a J.J.yYaa torepump an used. continuously during operation. With the needle valve leak adjuated fol" nol"nia.1 disoharge current, this pumping system will reaoh a pressure ot 2 s 10-4 mm .. ot mercury. A shut.off' valve between the osci.llogN.ph system and the pumps is provided. In addition, a. glow tube opel"'ated by a 7500 volt neon dgn t:ranstormer, a MoLeod gauge. and an extra adjust. able leak a.re oonneoted to the Tacuum 1ead in order to p?"Ovide means ot eontl"olling and oheoking the ?aouum in the system under any desired open.ting eond1t1ons. p?-eSSUN The glow tube giires an oeoasional flickering discharge at a cf 23 mm. ot mel"C\U'yf continuous diseh&rge begine at a pr&a$'\U't!I or 20 imn. o:f mercury and the tube becomes dark again at a.ppl'oxima.te17 . -2 2 x 10 mm. A reeees in the top ot t..'le lO'Weat pNselll"e stage of the di:ff'usion pmi,rs· aerNa as an adequate, although troublesome, liquid air trap requiring frequent re:tilling. sevenl ohangea wen made i.n the meohuiea.l layout of the equipment to 1.mproft th• ease ot opel"ation. The equipment wn..s originally set up tor- pNliminar-y testing on the south baloony 1n the Institute High Voltage Labontory.. contN>lt •re of' somm.at temporary conatruoticm. Exper1enoe ht thia f'ir1t 1ooa.tion indicated that atatio shielding of the oseillograph, c,~ a location more t-emote tr-om the surge genei-a.tor, or perhapa both, would. be dea1re.ble. Aocoriingly, a new looation in the aloove under the ma.in entrance to the building was selected beoa.uaa it would be more remote, and 11 if complete shielding were tound neoeaeary, the building itself would ao'b, a.s an ertec,tive shield 10 that minimwn. additional oonetruetion would be required. At the aame tb•, revis10l'l of the mounting and control appara.tu, 1f!il.S oompleted. Obaening hood.a w1 th dark sli4es were added abov-e the windows originally pNTided. The 10000 volt powr supply was ~ebuilt more oompaotly with the ti lament tranarorm.er tor the kenetron at ground potential and the high voltage tranatormer insulated for inveNe peak _voltage. A 5 milliampere meter, protected by a ael"ies resistor and a neon lamp, 'Mlt placed. in ael"iee with the 20 milliampere meter ht oi-der to give a more aeneitiw indication of discharge current. wire wo•d limiting reatators in the lead to the discharge tube repla.oed the we.tel" aolunm used prevlou•ly. An a. c. operated solenoid wa.s installed to di1..- charge the high voltage tilter condenser when the power was shut off. A self' ata?"ting aynohr-onous o1ook was installed to measure the oathode age, or length ot tm that a cathode had been in opeN.tion. Also, the high• voltage lead to the d1soharge tube was placed in a oopper tubing conduit in ol"der to shield the lead. The ooil _for the fO?"e-conoentntion :field•• r-nound. with larger wire so that a. single six volt ,torage batter-y would be su.fticient ae a poW:JJ- source.. A new compensating ooil ror the earth's ma.gnetio f'ield na wound, and the deflection bias o()ils were rnound to improve the unitomity or the magnetic tielcl. The resistances of the ftrioua ooila are tabulated belows Earth Coll 30.0 Detleotion Bias Coilt 7.. S Focusing Coil 9.,l Fore COl'lcentration Coil s.o 12 All of the oeoillog:raph equipment exoept the two storage battei-lea and the 1qooo volt poweli' aup>p1y is mounted in a dngle opeh.ti11g 'ba.oh. Below the osoillogn.pll are the d ..o pow;r •111pp1iea tor the 'bloclclng •y•tem, e11eep circuit bias, and detleotion oalibration voltage. A remoirable con• trel panel is oonneeted to a terminal blook (Tabl• I, tc11owing Figure S) by a cord and combination ot plugs so th• oontrole may be examined without in:bet':tering with the OpeKtion ot \he osoillogfaph, or removed tor cha.11ge• or ,epaira and ea.ally Npla.eed. The fteuum pumps are beneath the data tabl•. An electron ooupled oscillator with a seleoiio~ switch tor a ohoio• &t 100 ko, $$$ ke, 1000 ko, 8$33 ko, or 10000 kc 1e beside the oscillo• graph, with 1.ta pon:i- supply direotly be1ow. Figure 1 shows the geneJ!'a1 arNngement of the operating benoh. PHOTOGRAPHY Roll film with a rela.tinly slow fl1'111a1on, serusitivt only in the blue-violet l"eg1on ot the visible apec1:,um, proved to 'be most suoe•••tul with .this osoillog~ph.. Such film (Eastman NC roll fibi) 1i quite sensitive to dine-b eleotl"on expoaurei while at the same time it h relatively inexpen.sive, 11'. aan be handled easily because total darknel8 1a not required during development; it is not fogged it -the oamera ia opened when the open.tor's tao• against the vining hoed is the only ahielding trom the room light•• Film ori.1.nariily contains considera.ble moisture whieh DiUt be n ... moved before ,atta.taotoey opel"ation or the oeo111ogi-aph 1s possible. PhoaphOl'OUB pentoxide will dehydrate th• outer lay.,r-e ot paper •UffouruUng the film, but continued pumping with a Y'atruum aystem is the only adequa:te method ot preparing film for uae. A separate vacuum system fott this J>Qh pose would be deairable. Sinoe none was Mad.ily ava:U.able• a mambe>l" ot 13 f'i lms were placed in the oaoillogN.ph ll.t one time and were pumped oon• tinuoudy tor 24 houra or mor• until the films were dt"Y• Thctt JdNury pumps were not used until the glow tube •howed that the water Te.pol" pteasure had been Nduoed ad•quately. Dried .filmt were etor♦d in tht 0110illograph until they were requil"•cl. Two contraat emulaioas wre teated in the osoillogn.ph. Tu Nsults were a.imila.r in both oa.aes • the slO'Wlr •••P• gan an intense black tn.ee but the taste:r- ewop• wert· 'bar-•ly d:hcernible. Photoetat paper gave eontewhat bette,. l'elulti than proo.11a t'il.m., perhaps beo-.u,. this film. had a thin gelatine coating to protect the eulslon. Photonat paper might have some advantages it a large number ot oacillogi-ama an A :roll film. allows only 10 eeparate exposurea, to be made on one film. OI' so •urge osoi1lograms per roll it the r11m ii adTa.noed only on:e•half' turn between eltp0eun1. It it uaua.lly desirable -bo t!ffllop. the film. and .amine the re,ult1 after thia number ot exposUl"es has been made, and the ad.vantage ot a greater numtM;r of oaeill6gl"ama i,.r Nll becomes tuestione.ble. ) The hot that lere u1ature would be bl.ti-oduoed by a toll ot photosta.t paper would certainly be adTant•geoua. The photoatat paper might be n• movable in daylight if eome scheme 1imilu to that ueed with 35 mm. t!lm were adopted. This euggeation might be worth turther investigation. The etteot ot enulsion thiolm.e1s ma.y haw inf'1uenct4 theae N$ulta.• Emulsion• on roll films nry between thiolcwt1aea ot 1.6 ~ 10.-4 em. and 40 x 10""'4 om. The th1oknesa ot the photottat emulsion i e le•s than thee• ftlues but 'th• actual thiokneas •a not determined. The thiolmeis or the emuldon and. the prot•ot:lve coating on the prooees tilm 1iS 1.4 • 10'""' am. Woo4 9 ett.tea that the penetration or eleoti"'ons is propot'ttona.1 to the tounh powei"' ot the vc,1<,oity. The maximum penetration ia given bys 14 Ea• acoelera-b:1ng voltage d • denaity of medium. peaetnted 10000 volt el•otrone would thel"IJtOPe penetrate to a. depth of t.s • 10_. •• or leis tluui the total thiokneas ot the emle1on on a l"oll f'ibt. DEFLECTION D!STOR'tION D1nonioa .rteoting the nepoue or 'bhe beam to detleot111g fielde is ot majoi" im.ponanoe, u a true piotur'e of the elaotrioal tN.nsient being Noorded. is de&irable. cor"otions m1gh'b be applied to a. ditrtorted oscsillogl'am. but thia teemd.que ia definitely hthrloi" to ob..taining an oaoillogram wh.1oh NpN1ent1, within the toler-aru,• limit• of the exper1m.$!1t, a linear nlation betnen the measuN4 detlenion and the vo1ta.ge tnndent. A 11:nsu ••ep l• not always neoea,ary1 howev.P. ant1 an exponerttia.l 1-weep vo1tag• ia uaed tNqua.tly. Two methods or bias may be ueed to loo-ate the beam ott the aoreen. The beam may be dCJ.f'leoted to aey desind poai tion with a. :magnetie field.,. Ot" a fixed voltage may be applied to either, pail" of deflection plates. Conden,ere nmn be ueed to oouple the cletlection plates to the tranaieat volta.~ it th-e latter method. ot electroatatio bias 11 \UU»4. Originltlly, the twep plates wen oonneoted direet.ly to the._., eondenaei-a, and a magnetic tield was applied to retu:,n the beam to the acNOn. Two aet• or coils f'Um.ished tield.e to detleot the beam in the detleot1on dlnctlon and a.long tlw '"'"P a:d.1. These ooS.la wn•e 4 inehee in diameter and llitN plaoed about 4 inohet apart auproundil'lg the region of' the •-•P plates. The magnnio f'ield given by two ooill spaced a distanoe equal to their dlrunet•r is illuatn.ted by Fir;ur• s. The 1.ln•• ot equal magnetic ""o~ pnential •n found by graphio&1 inte%'p01at1a tl"Om 15 00000 ogooo 8088g 00000 i 00000 00~ oo ,,..... 00 88ogg I I I I I I I I §§ o o· 061"l0 000 . 8'6888 (/) ...J 0 u t... 0 (/) X <f. FIGURE 8 UNtS OF EQUAL VECTOR P0TEW11AL FOR TWO COILS SPACED TW\CE THEIR. RADIUS Figure 6 1 megacycle timing wave Magnetic Bias Figure 7 1 megacycle timing wave Electrostatic Bias computations ot '7.04( 4) Smythe 5 where Kand E are t.he complete elliptio integrals fQr the modulus k. k2 • 4a ~ [ ( a _ ~ ) 2 + 1 2] .1 a• radius of the loop ~=radial coordinate • • distanoe from plane ot loop As a result o:f this non-uniform field, a sine wave deflection voltage is reootded as a warped, double valued f'u.netion whoa• a.inp1itud0 ve.rles w:1.th position a.long the sweep rutis. {Figure 6). The oorreot a."'llplitude of the sine wave is given by Figure 7. A:tty attempt to obtain an e:xaet analysis of' ;uoh distortion would be impractieal, since the path of the beam at s:ny point dependi on the integrated effeet of a. magnetic and an electroatatie f'ield, an.d neither field is uniform, or even exa.etly known., Most of the detleet1on caused. by the e1eotrio field occurs near the top cf the sweep plateiJ the plates are not puallel but diverge to prevent the beam f'f'om approaching the plates during the extremes or the sweep,. since the beam is sv;l!'l)pt ts:t' beyond the left of the soreen in order to allow rebloeldng before the a'Weap returns. uniformly. The magnetic def'l.eotion along the sweep a.xis ooeura more S1.noe the magnetic field increases on either side of' the plane midway between two 005.11 separated by a distance equal to twiee their rradi'W:li the effect of the magnetio bias is greater as the beam is moved to either side or the zero axis by the deflection plates. the creaoent shape of the distortion. This fact explains The amplitude is reduced because the magnetic detleetion oecurs in that part of the field to the left ot the axis of the coils, where the field has a Ji'adial component in the 18 proper direction to account for the reduction in amplitude. Ccmpa.r-ison ot Figures 6 and 7 shows that there is a slight net ef'feot of' the radial component of the magnetic field 'P.hen the beam has a-wept to the rl~ht side of the axis or the ooils, i.e., the radial oom• ponents on opp&eite sides of the axis do not exactly oompenae.te because the beam striking the oenter of the osoillogram leaves the ma@:?l$tic and electric fields at the left of the eent:ral axis at sueh an angle that a continued straight path will inte:reeot the oenter of the oso:tllogl'am. Note also that there is an e111,tioal distortion of the beam spot when the beam departs from the tero def'leotion axis. Suoh elliptioal dia ... tortien of' the beam a.l•ys occurs in a non uniform field. In Figure 1 there h some slight crescent distortion at large amplitudes in the opposite direction to that previously discussed. This phenomenon is due to a. non uniform electric field oaueed by the iJhape of the s-weep plates and the ta.~t tha.t the sweep plates W9N not pb.oed .e ymmetrieally 1n the osoillogJ'a.ph. Th♦ extreme deflections exaggerate the e.f'i"ect. Detleotion distortion increases at a greater than linear rate aa the eleotrostatio and magnetic f'ialds are increased sim.ulta.neousiy with small resultant motion of the bea.m to the left of the soreen. It would be quite possible to reduce the etfeot satiaf'actorily by oapaoitively ooupling the sweep plates to the swep oondensel"s, and shunting the sweep plates wi-\:ih a high resiste.noe sc, the beam would normally begin its sweep at the center of the eoi-e-tm. A more uniform ma.gi'l.etio field from a Hebnholt1 arrangement ef ooils spaoed equal to their l"adius could then be used to bias the starting point of the sweep to the left. However, with capacitive ooupling it is possible to use an eleotrostatie bias whioh 1.8 easily oontl"Olled in magnitude with a potentiometer in the primary at the power supply tt"'anetermer. and the beam may be biased oom19 pletely to the left and the action of reblooking obsetted without tu intN>duotion of appreciable d!atonton. In oJ'd.er to provide a convenient means of ehanging 'the posii:iioa of the aero detleotion axis. the ooils for thi.a pul"pose wer-e !'&190un4 in a form vaguely resembling the int.raeotion or a ciroula.r oylit1der- and. a.n elliptical cylinder with a major axis twice the diameteP of the oirole and the minoJ" a.xi• of' the e111pe•. This form of coil gan a s.o mhhat more unt:torm magnetic field but some distortion Nmained,. although it ia - ' not !mportal'.lt when detleetions are small. BEAM DISTORTION several types of diatortla affctot the aile 01" shape of 'tM· fooue&d epot. Although this beam distortion i8 not quite as important; as detleotien distoriion, oenain type• of beam diatonion p?'()ved ve'ly dif:f'ieult to mil'limile. Fortunately• ho,ueTer, OJ>('Jl"ating eondi tiona can usually be oho•en •o that good osoillograms can be obtaiMd ove:r the entiN n.nge ot open.ting requirement,. one una?oidable torm of dietonia caused by a non unitoftl field has. bettn ea.lled elliptical distortion. conside:r- the effect or a:ri el♦otro• etatio field. u illustrated in Figu~e 9, on. a circular cathode beam ot gnater than intin1te8im.al diameter. If deflection li toward the un• grounded electl"ode, th$ eleotl"oas ne&l"est that eleoti-ode are aoted upon by a atronger field than that a.ot:lng upon the eleotl"one in the oppoalte aide or the 'beam. 'l'he i-eeult i• an olongatioa of the oiroular beam along the axis perpendioular to the eleot~ed••• ReduotiO!i ot the mi:nor axis of the elliptioa.1 beam will be oa.used by the curvature of the el•ctrio tield. Similarly, it the ungrounded e1eeti-ode itJ negatiTe, the oiroular bea.m will become elliptical with ita major axis parall•l 20 · F"IGURE 9 EQUIPOTt.NTIAL DIAGRAM NARROW PLATES ONE PLAT[ GROUNDED FIGURE WI DER 10 EQUIPOTE.NTIAL DIAGRAM PLATE.S ONE PLAT[ GROUNDE.D J - - -- -- - - - - - - - -- -- - - - - - - - - - - - , FIGURE II tQUIPOTENTIAL DIAGRAM ~ARROW PLATES PLATES UNGROUNDED to the electl"od.••. Focusing the beam would reduce the ef'feot tU.ghtly, ainoe the beam would. have converged somewhat before entering the dew fleotion field, but two focused linei appear tor ditteNnt value• ot Jocu.aing field current and a point tocue ca,mot be obtained. Th• end ef'feot.s introduce4 by the top and bottom of ilhe plates il1chaae this distortion oonsid.erably. Ceria.ln applioationa of the osoillograph, partioularly studies with a surge gen-erator, require that one of the defleoti~n plate• be g!'-ounded. A non uniform defleotion field is then una.void.able. While it is impossible to eeleot anr T&.ltte of tocusing field current mioh will give a naM"ow~f'aoe throughout the entire oeci1logtam., this distortion ean be minimiced by a pl"Ope!" ded.gn of' the detleotion plates. D•t1eot1.on amplitude is proportional to the product of the field strength and the time that the field acts upon an electron if the amplitude 1s ema.11. Increasing the length or the detlectio». plates will the:roe:fore re4uoe the :t1eld 1tnngth neoeausa.ry ror a given detleotion amplitude and deONase the el11ptioal dt•tonion by reducing the 1mportanee ot the end etfeots 1ntroduoed by th• 'bop and bottom or the pl.&tea. The \Uliformity or the .ti•ld ittelf oau be improved by oloaer spacing or the plate• and/or widening the plat••• Com.pan Figul'ea 9 and 10. All or these f'a.otora w1n"e con•ideNd in the d.edgn of new deflection plates tor' the esc111ogra.ph. Increasing the length ef the deflection platea and. J!'tt4uoing the spacing inoi-eate the oa.pacity ot the cletlection synem, which mAy not be important, and inona.se the detlectton aenait1vity ot the oaoi1lograph in addition to Nduc1ng 1:he e111ptioal distort.ion when on• plate 1, gNtmded. Mol'6 than 1000 -.olts -.a required foJ" a defie~tion ot •• in.eh i,rlor to changing the defleotion plates. 24 Detleotion aensitiir1ty was more than doubled by the ehange and it 'beoame pn.ctioal to mea.sU!"e detleotiont oon-esponding to only 100 volt$. TM.s increased eensitivity was q'11t. u•e:f'ul in tests of the blocking aystEnn and a pape.i-, "Icnization Time ot Thyrat1"0ns", contains aOJD.9 results of this inireatiga-t!on~ A preprint or this paper 1s included a.s Appendix A• .Anothet" pa.pe:r, "Surge Voltage Breakdo'Wi'l Charactel"istica to~ Electrical Gaps in Oil"• by Pro1'$•1or R. w. sor-ensen, ts iJt01.udtd as Appendix B. Thousands of aurge oocillograms •re ob'tained tor th!• in- veetlgation. TheH two pe.pera are examples of two widely dU'f'ere.nt type• of routine teat• that are possible with the osoillogn.ph. a aurge osoillogram and a auperlmpoaed timing ft.ft., FiguN 12 ahowe Now the g,.eate,- eharpness or the timing wave trace whioh ,vaa made with the two platcu, symmetrical with t"eapect to gl'OUnd. A better illustration ot t he efteot of eymmetr-ioa.l and unaymmetrioa.l detlection voltages oa.n be made by a compar-ieon or the surge traoe in Figun 12 and the muoh greater detleotiO:nl in FiguN 'T• .An annoying torm or beam distortion a.f'faoting the focus ot the beam during; n,eep, first observed in early te•ts of tht ~scillogl"tlph,, is illustrated in Figure 13. !f the beam is .tooused to a spot with the tube unblocked and. the tube is then open.ted without changing the focusing field ourrent, the t-raoe is $ha?'p at the b&gimd.ng 0£' the snep b~ th& traoe widtna almost uniformly dudng an interval or two micro•econds. ~erni detoeuaing has b••n applied to this phen.oD'lG,On. The The width or the traoe normally remains constant for the remainder of the s'W8ep after this two microsecond interval* although the width would occasionally decrease slightly after a maximum width had been reached. Usually the tocus was sharp at the exact beginning of the a-weep but in certain instances the 25 I\ I\ ,I\ A l'I A A A A Figure 12 Surge Oscillogra.m 1 megacycle timing wave l Figure 1~1 Transient gas focusing Interval of defocusing 2 microseconds beginning of the sv"8ep might be out of foous and the sharpest focus would then 00 1.~ ur during the interval of defocusing. This interval of' defocusing is independent of sweep speed; in fact, it is independent of every con• o~ivable variable that was checked during the three years of research on the problem. Actually. the term over.t'ocusing might be more descrtpt:i.ve of the phenomenon termed defocusing. In all oases the beam oan be focused sharply by reducing the current 1n the focusing field from the value required for focus of the spot when the beam is not blooked. This behavior suggests the possibility of tnnsient gas :f'ooueing which is not present duJ>ing steady state conditions. The correlation between defocusing and the use of fore-concentration was evident a.t almost the same time that the defocusing itself ·w-as discovered. The reason for such a oorrelation was not understood, howev-er, until the various factors a.ffeoting the distribution a.nd qua.ntibJ of stray electrons inside the tube had been investigated. Furthe~ basis for the mi~leading belief that fore-concentration is directly responsible for de. focusing vas furnished by the aooidental discovery that reve~sing the direction of the current in the tore-concentration field reduced the mag• nitude of the defocusing. Later tests tailed to v-erify this reduction of defocusing and this solitary example of reduced defocusing oan probably be explained by some oha.nge in the mechanical alignment of the cathode assembly while reversing the leads to the fore-eonoentration field. Very slight changes in the alignment of the cathode cause relatively great changes in the beam current. This behavior w-111 be described i.n detail in a later disoussion of beam efficiency. It is possible to obtain an electron image of the cathode surfaoe by adjusting the strength of the fore-concentration field. 27 "F;xruninatioa of the electron density distribution in the beam in this way ehows that moet of the electrons in the cathode beam originate at the bottom ot a. crater which fonns on the aurf'aoe of the oathode. An early obsern.tion showed that moving the image of the bottom of' the cathode o..ater out of the beam also reduced the defocusing. All of these first obsenations oan now be eXI)lained by the theo!'y or a oorrelation between defoousing and the presence of a large number ot stray oleotrons,, sinoe the ohangea involved 1nnriably reduced the number ef atn.y electrons by reducing the beam ourrein. The first intimation that defocusing is related to the presenM of a large number of strq eleotrone was given by the observation that defocusing is muoh greater than usual when the use or fore .... conoentration causes a tremendous reduction in the deflection by the eleotrostatio bias voltage applied to the sweep plates. J.i\U'ther investigation revealed that the current bet\\9en the plates shows a saturation ef':f'ect as an isolated voltage across the plates is inol"eased, indicating that the curNnt ic being limited by the number of stray eleotrons available. The satUl"ation current under the as.me voltage conditions was tripled when the beam current was neal"ly tripled by using fore-concentration. If the center tap of the voltage across the neep plates was grounded ... normal operating condition .. the stray cu:rrent without tore-eonoentl"ation •s increased somewhat and the use of fore-eonoentratien then caused the stray eleotron current to increase by a f"actol" of eigltt, The maximum measured stray ourrent caused a reduction ot the bias voltage of 10'/4 and the reduced deflection ot the beam when fore-concentration was used nrified this result experimentally. The reduction of deflection involved was muoh less than durin.g the previous expel"iment when the correlation between defocusing and reduced deflection with tore•ooncentration was first noted. 28 Grounding; the middle ephere or the trip gap S"ed.uoes the de.f oc\lsing. This eonneotion requires the application of the tripping impuls• to~ eutge recording to one of' the outside spher-es. Some inconvenience is oaused 'by' the neoessity of changing the connection for svges of different polarity but the system has been quite practical. The possibili~ that the entiN blocking system might be "tloating" at some potential above ground waa considered but discarded after tea~s showed that the distortion p~odueed when the blooking p1a.tes a.re Xta.ised to some potential above gr0und 1s quite odd ahapod and oan never be focused sharply, while photogl"aphs made with the sv.eep disconnected ahow that the defocused beam is oiroular !n shape. Encouraging ~esults were obtained by applying 90 volts to th~ blooking plates. A slight detleotion was produced and the foous was changed radially by approximately the proper amount and i n the coi-reot dil'ection to explain defocusing. Paradoxioa.lly, grounding the middle sphere increased the f'oousing effect, in oontrast to the opposite behavior :i.n the transient or true defocusing ease. However, when the vacuum ms improved by using a liquid air vapor trap, tests tailed to show s:ny .f'ooua effect with 90 volts applied to the blooking plates, although defocusing was still present during normal operation. of the osoillograph. Furth.et" teats without liquid air in the trap repeated the original reeults. The polarity of the voltage applied tc the blocking plates with t"espect to ground causes different visible transient effects but the steady state results are similar, a.lthollgh not exaetly the samo. No ohange in the defleotion sensiti~ity is apparent when the foous is changed by voltage on the blGoking plates. During tests ot b~am ourrent and stray electron currents as a :f"uncti on of pressure in the tube, a focusing effeot was noted when TOltage 29 was applied to the awep or deflection plates. This i"ooueing e:tteot 18 similar- to that obsened when lovr voltage is applied to the bloeldng plates. A saturation effeet can be noted, i.e., no additional focusing effect 1B observed ae the voltage is inoreased beyond ZOO volts .. Voltage was applied to the deflection plates with the beam unfooueeds the beam was then f'ocused to a point and the beam blocked and s•pt1 the Pesult was a b!'Oad traoe. The beam was focused to a point with no deflection volta.geJ then 'the same deflection voltage as used in the previous test was applied. and the spot beoame over-f'oousedJ the result was a sharp trace when the beam was blocked a.nd swept. These tests eeemed to prove that the toousing efteot obaened i~ the static case when voltage is applied and ■tray eleetf'Orui are preqent cannot explain the defocusing when the beam is blook:ed and thens-wept. Additional check of the theory that defocusing might be ea.used by the spark gt.p voltage a.oross the blocking plates was made by corm.toting batteries in series with the leads from the trip gap to the blocking platea. The gap voltage had been measured with the oscillograph. Visual ob1ena• tion indicated a total gap "10ltage or about 100 voltsJ photogn.phic measure- ment gave a nlue nearer 50 volts for the tO'ta.l gap volt.age.. Both 22½ 'f'olt and 45 volt batteries were oonneoted on each side ot the gap, in correot and incorrect polarity for oompenaating the gap wltage. with and without middle sphere ground♦d, yet no observable change in the detocueing ooul4 be noted. suoh definite negative r•aulte seem indisputable. A comparison of the def'oeusing when either lingle pair ot blooldng plates were used alone or when both pairs ·or plates were used, did not furnish any conclusive data. An attempt waa made to observe the traoe when the beam -was owept with the blocking plat•• diaeonn.eoted. sweep was obsouNd by the slower return tnveep. to be focused aruu-pl7. The ,-.,tun;. nieep appea.red Results 1Nre again inoonoluet-re. 30 The taet Defocusing definitely ooout1 and oa.n be photographed with the 1Wep p1•t•t' lie-. ool'IJleeted.. but with regular unblooting tiling p1a.oe. Thia tenlt pJtO"Nlt that the sweep t~d•nt is nc;t Ne,.nt!b1e, although it oannot 'b• •ald ·that the pre1enoe of' the •••P and/or det1ect1oa plates doea not, Wl1aett• the pheaomnon. Detoau11ng a1ao ocou:r-1 whh the det1eot1en p-lat•• U't :subetituted tor the regUlU' •••P plat••• Then is no change ia the detl•nlon ••ultivlty of the beam when the beam become• detooueed.. The d.etleotlon •~sit1v11'y -.s mtasund tor' beams of clittennt energies requiring tooud.ng ti.eld curNnt1 l.eee tlwl,. an.a greater thail; the toouling tield ournMs used tor tocudag a dt• f'oouaed. beam and an un'b1ooked 'beam. Then should have been a defienioa val"iation ot one.,fitth ot a.n inoh during the 4etoaustng intefi'n.11fi:th 500 vo1tt on th• detleotiott pla.tea it the etreot weN due to a ohuge ta the ener-gy of the eleotJ"Oq in. the beam a.bow the d•tle<ttion plat••• bn the -bn.oe •• a ,tn.ight 1ine. Detoouaing ia identical • n phot6paphed with the e1•otrou hitting an insulating gelatine emulslon, 6i' obaetyed vieully a• the e1ec:rtrona atrlb ·t l t1uoft1on.t ooating on an alum.in\ill plate. Any •xplaJ:lat1on baaed on deoeleft.'tiora ot the beam by an eleotf'oa "tilm" at the bottoa tt the tube oa.11 the:retoN be ♦llminat•d• The taot that the agnltud• ot th.• tooue ett•n l• •• gl"ft• • in n.re oa.aea a •harp trt.oe w1 th a detocueed beam ht.a 'b-,n obtaiile4 with ser-o magnetio tooudng f1e14 • amt the tact that gt-ounding the midd1♦ eph•r• ot the trip c•P lntluen••• the etteot, support a beliet that the tft111in\ tooull:ng OOl\tl"I in the uppeti l'egion ot the tul>e, pr-obably fa the 'f'ioinity ot the b1oold.ng p1atee., of rela.tS.ve1y high gas p,easUN. Thie part ot tlle tube 18 a ngln Additlcmal eviden~ It oontributed by 'bhe 'behavior- or the beam. when the tr!p gap i i not •~tnotl. with ,eapen to g!'Wnd.. Attar the trii, lead he.a been moved fliom the midclle fPhere to 31 an outside sphere, the b•am oa.m,.ot be tooaed at a.ll it iihtt ld4d1e 1pheN 11 \Ulgl"'etttn4•4s the trac• 11 a a11ea.r tor any value et tooua ti•lct au~. ;;--··~ : Gro\1l'lding the middle aphen thi-ough a reaifttanoe or several ~ t ! obml allows moat ot the tn.ce to be tocuaed. 1ea:rl.ng a "tan" at the beghmlng ot the •••P llhioh aamaot be toouaecl under any oonditlont. The tn.o hae acme amear- at the beginning or ••ep when the Yaeuum t• peor S.t the trip lead ia oonneoted to the middle apMN• A 1erlee ot teat• with vanoue trip gap comt.eotions led to the f'ollowbig oonolu•lont, it ftOUum is good, e! th.er d.ue to oontimiou puap. ing when the l"e1-t1ve haidl'by' of th• atmosphere ia low ao that tUm a.114 tube beOOJlle d.ey, 01' by th• UH ot a liquid air trap,. theft is little ...IU",i. ing of the traoe if the trip pp 11 0$!tplei:ely t1oa'ting, t.e., eailiffly diaconneoted. trom the trip hu. It the trip 1ea4 ia oonn:eotei to anr part ot the oiNNit, i\ tende \o maintain the eiroui\ at the oonditia before trippingJ tor exai,le, if the trip lead t• oonnmed to the 11t.d.dle sphere but tloating, it l• •••entia117 atgroum potential and the••• is aatiah.oto1"7 with a. good ••a. But, tt the trip lead 1• ooueetect to the middle •phere and to the tap between a pair ot unequal n•i1tor1 aoroH the eompleh gap, or it the tr-1.p l♦ad ie conneoted to an eutsiu a;h•l"e, the lead etteotiTely prevents the 'bloold.ng 1yatem from tJ•tnc sym:m.et:rieal with Napeot to gnund and th• Nault le a emeaNd ttaot. Th• tne• i• 1ati1taot<,ry, howwJO,, it the middle spheN la gnun4e4 through a 100 ellm ~•11-bor, t-egardl•H ot the nate ot the ftoua, t.nd the t..S.p lead can be oomieote4 to either 'the 111.cldle aphez-e or to an o\tteide sphen. O,ie othel" obtern.tioa should be unt1one4. A tepo.....,•· rleibly radial chang• i». the ,11• ot the toCNeecl 8P-' oe ba o'bael"ftcl oota1!onall7 at t~ e1eotroata.t1c bias voltage 1a applied to the sweep plate,. No in. veati.cation w&8 mad• of the phaamenon, as ite ooounenoe •• quite l"ar'ea Th• phenomenon may be explained by a tnnaient ga1 f'oeu•ing caused t,y morlng the e1eetnn beam into a runr pa.th. deatl'Oying the equ11i'brlua be- tween the poaltivo ud negative space oharge1 in the t'egf.on near the eathode beaa. On.e tactor- Nma.ined imponant thnughout the entire lnveatigati.011. Any b1oree.se 1n the :num.bel" or stray eleotron.1 inoreases the ctetooudng et.feet if' other taotort NJD&in ccmsi.ant. Poor- ftouum invariably oau1•e detoousing. but no d.if'terenoe ·oaa be detected behven a pool" vaouum. du., to ,va,ter Yapor within the iNlM· os- a ds-y tube with the a.1r p,eliure equal to the partial pree~un ot the wateF -.pa 1:n. 1;he tint·.;~a.••• Stf'ar eleotrona inorea.se a.1 'the b.eam cul"Nnt in~reaaes~ regaPdlese of' the oaue• of the inoreaaed b.am ovffnt, but not neoe•tarlly in p:report1on to tt. This behavtor -.plaint the oorrrela.tion bet-we.n d.e:t'oo.ueing and fore ... oonoentn:tton. No detinit& relation could be disoOTePed. between d.ef'ocusing and the JiUlllber ot atfty eleotN>na pneent at d.1tterent looa:b.ions ,T.i:th!a 1:he tube. In. taot, def'ooueing ,.. lmown to ooour when sti-at ele·e trona were l••• .num.eroue than. dving a pnvioua test in which detoaustng did not ooffil!'• Gail tocn1sing alone Oant\Otbe conaidend an adequate e,cpluat:lon. Gas tbau.dng would prediot an untoouaed beam at the begln.uing of, IS1NMp;i with the net ot the trace tccuied af'tcfr ionisation had oeeurred, du• w the telease ot tihe 'beam. Th:la behavior does not 0aov. some alight gas tooueing •s observed in tM 1tead7 state with the beam not blocked 'Wh.c the beam OUJ"J"tnrt • • «roeptiona.,lly high; i.e.,, gNater .than 60 mioroailpff'H. Aa fvidooe tor the existence of •ueh gas tocus!nr;, the trace was undGJI'• roouted at the start of' the._.,, pe.ased th..ough a point f'oous1 then noa.me badly detoouaed. A oomplet• es:planation wb.1oh sktiaf'iea all of th• obsettationt relating to def'oousing would be quite complex. can 'be ba.ced on a tN.nsient gas f'oousing. A simplifi•d. explatta:hioa The interval of defoeudng la the time i-equired £or the formation of a positiYe ion space charge. Normally; the ions ma1nta.ln the gas tocueing .for a pe-rtod not longer ih~ the perailt•nce of vision. The duration of detoculing with the b..._ stationary wa.e oheoked by U'abloeking the beam by shorting the blooldng plates af'ter the blocking 'f'olts.ge had been d.iacoJUSeoted.. The sweep ba.. pulse was not applied to the 8?,1eep plates during this test. The ep l)eJII• oeived an over:f'oouaed spot whioh changed insta.ntly to the point tomp., oharaoteristio of the fooused beam with the beu not blo~ked. Perhap• the dafoousing perdste as long as the beam h moving into a region ot lUl• ionised gaa molecules. At any rate., some relation b&tween the mobilitie• of the ions and the stl"l.y eleottom ores.tad whft the ga.a moleeuletJ beo•• ionized allows the neut~ali•ation of the apaae charge oauaing gas teoueing .when equilibrium has beem eatablit'lhed, but permits the ooouJTenee ot a transient ga.s toeusing.. The number et positive tone would 1.norease aa etny •lectrone ine:reased, but would not neoeaoarU.y be 1n the foi"Jll ot a space charge whioh would cau1e gadJ tocusing. suoh an explanation doee not offer a good correlation with tldJ behav:tor l'lhen the mid.d.le sphere of the trip gap 18 gl"ounded6 But moat other <ib•enations can be explained, at least pmially, b7 the h7Pothedt. The uniformly l"adia.l roousing etfec'b ot an eleotroete.tic fi•ld may be explained b1 a. steady state gal tocuaing productd by the eff•trts or field on the motion or the poaitift ions oi" "bhe atn.y eleotrons. oi- bQth, pre~•nt1ng the normal equilibrium. hom being eatablished under etea47 ata.te oondit1ou. It is knom that electrio :f'ielda 11.ff'ect. the opvatt• or hot-cathode. gae.rooueed cathode-ray tubes. 10 J.B. Johnson mentione a momenta1"Y "pa.use" in the tnoe ot a sine nve in a. tube using ga.t focusing• c8.used by the neut:r'aliza.ti on of weak ele ctric fields by the space <>harge. This "pause" when the deflecting field was weak W(I.S never obee:rved with this oseillogre.phAt the pt"EHient time there is no 1.nforma.tion to prov-e that thili defocusing behavior is not limited to this particular- osoillo~raph.. However, it is likely that the phenomena h:3.ve been observed by others. Fortunately, de f oous l n g can be minimized by using a l ow b~am... current f'or the e1•ieep gpeeds Go i nciding with th$ int erv1tl of defocusing. The trace wi th a.."'l unconoentra.ted berun is invari ably sharper and preferable to a trace using fore-concentration. Fore-concentration i s prilnarily useful for increasing the visible :lntensity of the osolllogrron when o'b• serving w:Y.th a. fluoreeoent so:reen, or for photography with sweep spaoda approaohing the ,~ l ting speed. limit. In the latter ease_ t he interval of.' defocusing is very long oompa:red to the period of $"Weep and defocusing is not objectionable. Good oscillogre.ms a.re therefore poss:i.bla throughout the entire operating range of the oseillograph, if' careful technique is employed. A third form of beam distortion was observed occas i onally \vhen an abnormally large amount of water vapor was present in the vacuum system. The observed af'teot ·was e. series of unfocused beads connected by a tha.rply focused trace, Th1.s bead distortion is illustrated by Fl f,'i re 14. (The effeot became familiarly known as the gopher snake beoause its appearance resembles a well~fed snake of that species). Bead distortion always ooou:rs d.multaneously with an abrupt de• orea.se i .n the di$oha.rge-curren'\;, other factors remaining unchanged. 'I'hia decrease in diacbarge•ourren:t represents a discontinuity i n the diaohargeourrent ... discharge preuure oune,. 35 The air pressure in the d1.scha.re;e tube 100 Figure 14 Bead Distortion Time scale microseconds 5---------------------..--------------------, Ill Ill a:: -4 w Q. 2 ; 4: J -I ; / / 3 ; ·i ; ,; ... ; ~ 'Ml a °' '2 ':) V w .I J er < ::i:: ,v (II 0 O'-----....c._...:.._ _---1._ _ _ _ _ _ _ _...L._ _ _ _ _ _ _ _' - - - - - - - - - - - ' N£EDL.E VALVE SETT ING J:.IGURE 15 was not known, but it is not unreasonable to believe that a oontinuous change in the setting of the needle valve leak gives a oontinuous va.rh..ti on of the pressure within the discharge tube. Figure 15 shows the ~e- lation between disoharg;e-ourrent and needle valve setting during the period when bead distortion exists. Disoha?"ge-current begi.ns and increases to lqO milliamperes as the needle ~alve is opened. is sharp throughout this first region. The osoillograph traoe No defocusing is apparent. Then the discharge-current will drop suddenly to 0.5 milliamperes and bead• ,'ti ll appear in the trace occasionally. With the beam unblocked, the beads appear as a "halo" around the focused spot. The unfocused spot is not as sharply defined as it would be with a. normal discharge~ Usually, the d:i.scharge-current creeps upward to a normal value and the beads disappear if the needle valve leak is inoreased until the disoharge-ourrent becomes 2.5 milliamperes. Figure 15. This is illustrated by the vertical solid line in Ocoasiona.lly, hoW:tver, the change from a low disohe.rge-eurrent to a higher discharge-current is disoontinuoua. Decreasing the needle valve leak causes the discharge-current to decrease to a value somewhat less than 2.5 milliamperesa then a discontinuous drop of approximately 0.5 milliamperes occurs and the disoharge-ourreirt remains low until the discharge disappears, due to an insufficient air leak. The behavior as the needle valve leak is decreased is shown by the dotted line in Figure 15. The duration of the individual beads is invariable, although the frequenoy of occurrence of the beads increases as the disoharge-eurrent is inoreased from 0.5 to 2.5 milliamperes by opening the needle valve leak. I~i gure 14 shows three beads during an interval of 200 mioroeeoond•• fr$quency of ooourrence is seldom greater than this value. The On one ocoaeion the frequency of ooeurrenoe was approximately once in 500 to 1000 mioro37 -seoonds with a discharge-current of 0.5 milliamperes J inoreasing the <lit• oharge-eurrent to o.8 milliamperes inc:-ea.&ed the f'requeney or oocurnnce and the beads appeared oons1stently in sert1i-eonneoted pairs. Further in• crease of the discharge-current to 1.0 milliamperes ca.used. semi-connene4 triplets aep8.1"'ated by a relatively sharp trao•. These multiple beada are not the usual phenomenon; ho,uever. Bead d.iatortion and the :region of low disoharge-oul"l"ent disap~al" immediately when a liquid air trap is used if the water vapoJ" w.l thin th• But continuous pumping for aevenl tube has been introduoe4 by new films. days was required to reduce bead. distortion when the water vapor 'Wal in• troduoed by a. wet towel wn.pped around the needle 'ftl.ve leak. Two independent cheoks proved that the beads must be a focus ettect and cannot 'be caused by periodic ohanges or &coele:tation voltage due to a variable dl"Op across the 11.mi:bing resistance in the lead to the cathode. as the diaohar~-oul"l"ent tluotuates between the two Ngions ot CU.l"Nnt indicated by Figuh 16. Changing the value of the limiting ttesistanoe produoed no 0bservable difference in the si&e, shape, or the hequenoy ot occurrence of the bead$. Voltage was applied to the defleotion plates while bead distortion was oooutting. No change in the deflection s$!Uiltivity of the beam occurred when the beau appeared, a further 1ndioa.tion that the effect is oa.used by a f'ooua phenomenon. An interesting but uncottelatetl observation is \hat no change in the voltage acrOS$ the t2:-ip gap ocoUJ"a when a bead appears. This obsel"T8.tion wa111 ma.de with 'the bleeldng gap oon- neoted to the detleotion plates. There is no obviou1 explanation tor the simulta.neou.e ocouttenC(t of a periodic focus change and the disoontlnuoui beha.~1o~ or the discharge• 11 current. M. von Ardenne describe• an anomaly termed • Ionensohwin.gungen' whioh gave a periodic focus variation, but the cathode-ray tube was a 38 .. Some data by G~h•~ohulwse sealed glau. hot-cathode. gas•focueed tube. quoted by Loeb 6 . . . .. . - show a non linear variation of the cathode tall ot potential as the percentage of hydrogen in a mutur• of mercury and hydrogen is varied from O to 1~ hydrog•n. A similar behavior might explain the disoontinuoue behaviot> or the discharge.current in a mixture of air and water vapoi". In d1.seussing the anomalous behaviol" of discharges in a mixture of gaee,, Loeb comments, "...... . . . . . epeoule.tion is futile .. " ANOMALIES Se-reral inter-esting anomalies ooour~ed. Most of them were un. important, but some 1¥81"9 sufficiently OUl"ious to be worthy of men~ion. At one time a taint glow appeared on the fluoreeoent toreEtn when the beam was biased to the letb of the screen and blooked. This ghost bee.rn behaved quite normally tn a magnetics field_. moving 1.n the legitimate direction. But as the electrostatic bias was decreased and the beam ff• turned to the eenter of' the screen, although blocked and not visible, the ghost berun disappeared in the opposite direction. The ghost 'beam is r-eadily explained by a tew stray electrons ot high energy whioh p&.H$d through the blocking windows 8.l'ld between the le:f't sv..eep plate and the b:nsa vmll ot the oscillogN.ph tube. The electrostatic f'ield between the plate and the wa.11 ot the tube ha.a a direction and magnitude which deflected the stray beam baok to'WU'd the soreen. The ghost beam disappeared when the oathode then in use wae replaced.. Another rather a.mating effect was observed during some tests with a single shee-t of process film in the osoillograph. Fi.gun 16 1a a photogn.ph made by unblocking the unoonoentrated beain with sweep No. 6 1 the slowest available ,weep spe"-, dieconnected from the sweep plat••• The beam is split in-to two eecttiona which have separated t11ghtly. 39 The j J Figure 16 Figure 17 Figure 18 dagrae of "splitting" of the un.fooused beam deoi-eased. u the unblocked i nterval was shoJ'tened by selecting raster swM:p speeds. When fil"at Un• blocked by removing the voltage supply, the beam had a norm.al oiroulu shape w:t th a te.int line through the beam a.a i.n Figure 18., The b♦ am would then fly apart and gradually attain an appearance simila:r to Fi gure 17. The figures ha.v-e been drawn to the same scale as the photograph. Laak of additional tilm in the oamera prevented taking other phetographs of t he phenomenon. The :film -was then Nmffed., time e.nd inspected. The oath~de wa.s nmoved. at the same It had a nol"m.al appearance with a barely formed crates-. The same cathode was N-installed without part:l.eular ea.re to obtain the previous alignment. (This teat ooc'Ul"red before the significance of cathode position was diseovel"ed). The beam was at first satid'a.~tory, bU.t still had the shadow line shown in Figuroa 18. An hour later, with the 'beam •tt during the interval, the split beam returned. The air leak at the needle valve was aooidenta.lly out oft ancl the beam became oir-oular again, with the As yet, no fore ... oonoentratian had been shadow position. changed somewhat. uaEtd. A w-eak ro:re-conoentration field widened the separation of the beam at ·the shadow.. The position of the ehadow did not ohange as the eletJtron lmag$ of the oa.thode we.a moved to dittet"ent ngion1 of the beam by ohanging t he position of the f'o:re-oonoen'bration field. One e?-itieal value of foN• ooncentration field-our:rent and poaition caused. the beam t o tly a.part. The pattern differed tr-om Fig\U"e 17. This dif'f'er-enoe wa.e expeoted, since the position o.f the shadow had changed as noted above. The split beam oould. not be focused to a point; and the bemn did not appear spli t if overfoeussd. Eventually., the phenomenon disappeared. It is dli'ficult to attempt to explain a phenomenon at a later date when nevr taot&rs are dieeovend., but the behavior as the Ol"'iginal cathode 41 aged (a new orater. "Weuld undoubtedly be tol"med af'ter the oath<,de 'W&8 i".. installed, and the e:t'tect would be identical to the installation ot a new oathode) could have been oa.used by the first audden decreaae ot the beam• ourrent me.king fore-oonoentHtlon neoeeeaey to maintain the intensity ot th.e beam. A minute obetNotion. which was eventually burned away by th• beam, must have ca.used the obeerred shadow in the beC11. Probably the obstruction was located at the limiting aperture below tha a.node 1\ri.ndow, as its position changed when the needle valve leak was accidentally eut off and disappeared soen af'ter fore-eon~entration was used. On one occasion, the unfoouced bea:m. appeared as two distinct oon• centric beams, one larger than the other. This beam could be focused to a spot with a sur?'Ounding undert'ooueed or overfoeused beam by focus field ... ourrent values ot 0.290 amperes and 0.315 amperes respectively. current was the euri-ent normally required for focusing. The higher An attempt to measure the deflection sensitivity of the double beam failed beea.use the dual oha.raoter of the beam disappeared when def'leotion voltage was applied and slowly returned ,vb.en the deflection voltage wa.s removed. The double bea.m did not seem to be affected if blocking voltage was applied and immediately removed. Continued bloeking of' the beam oa.used the beam ~o lose its dual oharaeter although the double beam :returned slowly if the beam was lef't unblocked. A list of the varioue anomalies, thEiir oa.use and correction, lt known, is given in Table lI. 42 TABLE II LI ST OF ANOMALIES CAUSE CORRECTION Bead Distortion - Periodic Water Vapor in Rem.oita.1 of Overfoousing Phenomenon Vacuum System Wa.te,:t Vapor St r ay beam with blocking Old. cathed.e New Cathode Ob&1tl"U.otion ObstJ!'!tetion between anode disappea.Nd ANOMALY voltage applied Berun split and separated and bloeking oh9.l'llb&I' Two e0neentric beams Unknown Oocurl"'ed mily onc«h El im.1.nated by eleoti-o• statio fteld Cyolio variations t>t See discussion Aging a new Beam Eff'ie!.enq in BE.AM STUDIES cathode 43 ELECTR01'T OPTICS The use or ft. cold-cathode oscillograph as an eleet:rQn microacope has considerable praotieal value as well as theoretical inte:vest. Buso1.i1 has shown that a short e..xial f'oowd.ng coil acts like an optical lene. axial mat~netio fields are ava.:i.lable on the oseillogra.ph. 2 T\'UO One is normally used to focus the disoharge on the anode, window and inerease t he baatn eu:,... rent; it is lmO'Wl'l as the fore~concentration field. The other is used to focus the cathode•bea..m to a spot on 'the fluorescent screen and will be N• rerred to as the focus field. These fields a.re given by the leakage flux tr0m the air gap in an iron oa.ee with axial symmetry. 1'he a.na.legy to o. thin lens is therefoi-e quite good,. An image of the cathode surface is formed on the fluoregoent sereen if the fore-concentration f ield is adjusted to a value somewhat less than that required f'or maximum increase of the beoon current. The f:i.eld of the image is limited by an aperture located bet·ween the anode and the blocking syet$m. Thie aperture is required for operation of the tube a.s an oscillograph. With t:ero focus field ourrentg the image is a.bout 1.2 cm. in diameter and represents a magnification of approximately 50 diameters. A f'ocus field current of 0.30 amperes is normally required to give a point foous. or image o:t' the anode window a.t the fluorescent screen. The distanoe from the anode window to the focusing field ie 45 cm. pd the distance ft-om the focusing field to the fluorescent screen ia 55 om • .Increasing the focusing field our-rent &om 0.30 amperes to 0. 35 ampere• decreases the image distance f'rom 55 cm. to 35 cm. I£ full battery voltage ( 6 volts) is applied. to 'the toeue field, the field eur!'ent will be o.s, amperes. Thia value or tooua field current forms an image of the anode window wel 1 a.beve the :f'luot'esoent sereen, and a fllpot t,,vi.oe as l a rg$ ae -the unfocused spot a.ppea.rt on the fluo?"'esce11t aorep. Since the •1•••..., from a pattieular point on the os.thode •l.U"t•ee are tn•eling in pathl easen:bially pai-a1le1, beoauee of the small diameter of the anode. wb:ld.ow,.. the focus field: pr"Oduoea little ohangfJ in the foous of' the electron~• of the cathode au~taoe, and an enlal"ged but inverted 1.Ju.ge of th& cathode -app$&rs on the tluoN•tent screen. Addl1d.onal voltage tan b$· a:ppl!ect ,o the foous field, inOJll'eaaing the taa.pifioation to ai)pi-oxbis.te1y 500 diamete:,1, but sharpness er detail i, saoritioed. duced by both axial fie hie. Some ~tton of the bulge is no- The rotation seldom ezo•d• 90 d•CN••• Thia arrangement or a pUJho1e aperture and t . magnetio "·oonvet-giag lenses", one to oonti-ol the f'ooue ot the image &ltd the othe~ to vary the· maviif1oat1on, dlttere ttf)m. the el•o1.l'on mioro•ooptt• deetribe4 by ZWOJ"Yldn11 , 8 Rusm14 , M.yeri 7 • and Ma1otf and Epate-in , but the design 11 quite eft.otiff. sev•ral optical analogies poasibla with th• aysttlm. are 111u1ti"ated by FiguN 19. Figure l9o ii the ei:ilplest analogy, ud has been eheoked with a tlaeh• light, a pinhole, and a oamn. lene. The .image from auoh an arrattgeaent is, of oaurse, poor-, due to the l.a.rige magnlfioatia. fl-om a pinhole leu. It ts po,a1ble to enmin& the el•otMn image or the ea.thode suf'ta.fle and obeel"Ve the cha11gee which oeou,, a• the oathod• 1s uaed. This tact wat usefUl in ob$eni.ng the roi,nation of a ohter in the end ot the ••thocl•• and showed definitely that the sudden d.&oNaile in beam etfioienoy art•~ a cathode nae been in uee tor app~oxbaately 5 mimate• oolncld•• with 'bhe formation of suoh a on.tez,. (see Figutte 2S tor Ci\lttel ~ bMm oUtteat &a a function of oathode a.g•). The image ot a new oathOde thows that the aouroe of the •leetl"ona i• Wlunoed by lathe tool maJ"ka a.n4 ·stmt.1u av. t"aoe il'Te-gulariti•• • Th• actl.011 ot poeit1..-e ions attiking the •athod.• surfaee eoon nmov•• theae :t:,regula.rltiee. appears. A aresoent-lha.ped bright epot This spot ino:ree.ses in aNe.and torma a ,emi-oi.Pol• w.tth a. bi-igh1 45 CATHODE APERTU.:.:.,RE:::.__ \ \ i I I i . i \\\; /I \I '.\I CATHODE IMAGE WITH FOR£. - CONCENTRAT'I ON a. FIGURE. l9 ANODE i IMAGE MAGNIFIE:0 CATHODE \MAGE NO FORE. - CONCENTRATION b OPTICAL ANALOGIES FOR lHE C · OSCILLOGRAPH spot in the middle e:t the ara, like a marble testing in the bottom or a bowl. The arc gi"adually beoom.es a oan.plete ring, and the bright apois 11tCWea to the center of the ring with a Nlatively dark area between the central spot and the outer ring. weaker ring. Very old craters may show a. Hoond larger ati4 The sise of the ring va.riea ,tlth th8 age and she <>f the crater, or oou~ee. Experience has shown that .rore-ooneentr-at1on is not a.a va.luabl• as might be expeo'bed from the increased beam--eurl'ent '?mich it makes a.va.11• able. The traoe oe.nnot o• focused aa sharply as the trace vrl.thout fore- ooncentration1 as a. result, the width of the t.n.ce increases and th• apparent blaokneas of the trace is not increased. Dlff'ioulties due 'bo tt'ansient gas focusing a.re in-oreased by the increased beam-o'UJ'rent. FoP these rsasons, tore-eonoentration has not proved satiatactoey fo~ photogi-aphio recording excopt at the axtNmely high smep speeds. FoM• oonoen:tration is quite valuable for visual obsc,natione, h.o'W8Yer, as th• bJP1lllance or the t:n.oe on the tluoresoent sol"tan is inereased considerably by the incNaeed beam-current. Fore-oonoentntlon increase& the beam•ourrent t .o a value six time• as great ae the ftlue without i"ore ... concentt"ation if the total discharge. is small, i.e., o.s milliampere,. The efficiency or fore-aonoentN.t1021 de.crease• ae the total discharge our:rent is increased, but the b4Jam• cu:rreat inoHasee 'lltttil an optimum. value of total dieoharge-ott't't'$Itt 11 rea.oh.i. Furthei" in~rease ot the total d.isoharge cuttent ea.uses only a. slight reduction of the beam-current if' no fore-ooneent:ratton f1eld ts used, but the ettioienoy of ro:re.• eoncen:tntion beoomes quite low after the optimm value or ,otal disoharge ourrent is exceeded. Note the curves of beam- ourrent as a function of fore-oonoett:tration f'i&ld--outtent f.or dilcharg&• 4'1 CUl"J"f)ntS or 4,.0 and 5.0 milliornperes (Figve 20). The values of 'b....►. ou.rNnt with.out toN.eoneentn:bicm, and baa:m..oui"Nlnt when the :toi-e.. eonoentn.t!on field ourrtnt is adjusted to its optbwm value, h~,re 'be• replotted in Figure 21 as a f'unet!on or total diaoharge...m:rr-o:n'\f" Max!.m:mn beam... curt'ent oeours wh.$11 the f<H•e-eoncentra.ti(m t1el4• current is adjusted to a ,ralue $0lllfMla.t highel" than the Ta.lue r~il'ed for an electron image ot the oathode surface at the tluol"eaoent se-r-een. lf. thel"e were no limiting aperture below the anode, a JJUAXimum ahould 040V' when the discharge is fooused on the anode window. The p:reeenoe of th♦ aperture requires that the image of the oathode be tocused at a pl..- bet'i/18en the a.node window and the aperture. Influen.oe or the magnetlo field upO'i'l the na.tur~ of' t:he d:hoharge ~ at.ten the a.djustm.&1tt fff maximum beam-ouri"ent. particularly when the total diechar'ge-oUJtrent t.• high; althoue;h the fa.ct that the maximum beam-current occu_nied whe the ove:rf'ocused image of the bottom of' the cathode crater juat .filled th• field. limited by the aperture, indicates t:ha.t the geometrical faotor-1 conti-ol the maxinum. beam,...ourrent for any oonstant, low value of diaohar-g.. current. A btu.\m-ourrent· as high as St of the total dieoharge-ourrent has been obtained with low d.iecharge ... current (o.3 milliampere•)• Beam et• fielenoi E3S of 5% W9N obtaintd eonstatently with d1ceha.Pge-o'U.J'l'N:nt n.lues of o.s milliamperes. Disoha.rge-ourrent 1n.cre:a$el a.s the fore-eoneent:ration field strength is increased if' no adjustment of the ne•dl$ ,rab-e leak is made to l"e.d uo• the pressure in the discharge tube. Th1* oUl"rent inorease continues be- yond the -n.lue ct fol"e ...oonoe:ntration f'1eld-current whioh gives the maximum beam-current. As an example, the total dis()harge•out'rent was adjueted to 1.0 milliampere w.tth aero fore..oonoentration field ..QuF.rent .. 48 The beam- 90.------,,-------,------,,----------------------, FIGURE 20 ,BEAM CURRENT - FORE CONCENTRATION i A - 0 S MILS DISCHARGE CURRENT 1B _Jc BOt--------+------+------· - IO MILS DISCHARGE CURRENT 2.0 MILS DISCHARGE CURR£Nl D - 3.0 MILS E - 4 . 0 MILS F - 5 .0 MILS iI i i ! i I 701----- -- -+--------- .. i I ~601----- - - - - 1 1~ . : .. - ---· -· •. - -,·-·· - ·•···-- -- --- ---- . f - - - i J ~ f- z L.J Ck:' 40-1 ----- j _ _,__. _,__ - - - er .::) u ' I I I ... ·~,--- - -·· - -- - --- --!------ - ·- · II 20 10 l 0 .25 • .SO .75 I 00 1.25 150 FO~E C0NCENl'RATION FIELD CURRENT- AMPERES 1.75 90--------------------------------------:i FIGURE 21 80 1 - - -- -- - L -- - -· - BEAM CURRENT- DISCHARGE CURRENT A - NO FORE - (ONCE NTRP.TI ON Fl f LO __ _ __ __ _ B_ -__ o_PT_I_M_1J_M_A_DJ.,......u_ s_ r_M_E_111_ro_F_FO_R_E.,......·_c_o111_c_£_N_1_R A_r_,o_N_ i I 70 - - - - - - - t----- -- - I' 60 -- - - - _J_ _ _......_ __ _____ - · - ),- z lJ.J C( cc ::> lJ JO 1 - - -- - ~ <t uJ co 0 1.0 2.0 3.0 DISCHARGE CURRENT - MILLI AM PERES 4.0 5.0 TABLE III FORE-CONCENTRATION FIELD-CURRENT VALUES FOR IMAGE OF CATHODE A.JI.TD MAXIMUM BEAM•CURRENT Total Disobal"ge Current. Milliamperes 0.5 1.0 2.0 :s.o 4.0 s.o Beam Current with Zero Fore-ooneentration Field. Miaroamperee 4 12 2S 21 ZI 16 Mu:im\ml Beam Current. Microsmperes 24 48 76 iS 24 1.00 .90 .as .so ,.16· .,o .es .75 ."12 .'10 .65 .es Fore.oonoentration Field-Current tor Muinmm Beam...current. .Ampel"eS Fore-concentration Field-Current for Qa.thode Image. Ampert!!e 51 current under these conditions was 10 mioroamperea. When the current ta the :f'ore-oonoe11.tra.tion field 'Was increased to l.O ampere, the total dis• charge-ourrent increased to 2.0 milliamperes and the bea.m.-ourrent to 60 mioroamperea. Increasing the tore-ooncentntion tield-nur-rent to l.S amps!'es increased the total disoharge-cvrent to nearly 3 milliamperes and reduced the beam-current to 12 mioroanrperes. This 'behavior ot the total discharge-current ia readily explained by the action of an axial magnetic field on a glow discharge. The data plotted in Figure 20 were obtained by readjusting the needle ...alve leak so that total disoha.rge-oul"rent was constant at the Talue shown for Moh curYe. No adjustment of the aooeltPating voltage was ma.det each cune in Figure 20 represents a constant voltage but tho TOltages fo~ the different oul"Tes differ by the drop aeross the limit~ ing resistor'. A well-aged oathode was uaed in order to avoid f'luotuatione of beam ef'tioienoy (see FigU!'e 23), but aome variation did ooour during the test, which required more than one hour. and it was not al-.ys possible to repeat values. For this reason. Figures 20 and 21 are merely repre- sentative data, and certain unavoidaQle but relatively unimportant va!"iablee are include4. A list of additional data not shown by the curves is in• oluded in Table III following Figure 21. BEAM STUDIES Beam eff"icienoy is in:f'luenoed by a number of othel" raotora as well as by concentration of the discharge with an axial magnetic f'ield. The etteot or water vapor within the tube we.a mentioned in the dilcu$Sion ot tnnsient gas focusing and other forms ot beam dittortion. The increased numbeit of atra.y electrons whioh ia always associated with a. poor vaou'llll must be aeoompanied by a decrease in the beam-eurrent, since the atrq 52 eleottou are tormed. by the oollia!.cm. oi" intera.ction bttnen ca• mol♦oui•• and the e1•otront 1a 'h• beam. Figute 22 illustra.tea typical d.ata whioh show the improvement in the beam-cu:rNnt a.a wa.tez- npo:r is f'•OV'N tton the •yatem by- a lif4Uid a.ii- tn.p. Wh•n liquid air was t111wt addM to the trap 1 the beam-cUl"Nnt wa.a 12 nd.oi"Oa:mpene., and the athy e1eGilffn.- ourten'b at the right ewep plate, ua.sured with the positi:ve teminal ot a 90-volt battery oonneoted to the adep plate and a mier-Oi.ln:metel" bl 1erle1 with the lead to ground. wa.a 18 m:!.oromaperes. Beam-wrrent in- crea1ed to SO 111.ol"Oaapel'e• ,m.d the ltt-&y eleot:ron-ounent deoreased to ti- mioroamperes over a. -period or thr-ee hour,, and after tb&.t tlme rema.1:ned. steady. The length or time required f'or complete remon.1 ot •t•r vapol" f'rom the eyetem ns not al1nL7s as ghat aa thtff hoUl"s. In some oa.see im• provement was almost instantaneous. Stray eleotrona inona.ee m.oN Ji'apidly than the b•im•CUPNnt 1'Ml1 the beam.current ia increased by a tote•oonoentntlon field it the ftOUUll is poor. But the Patio of beem...cun-ent to the ,tray •l•otr•-cvl'"em: at the right .,..., plate inoreaaea as b•am.-oUM"ent !.a inanased when the oaoillognph 1, f'Ne ot wa.ter vapor and the aiio pnesuH ia 2 11: 10°"' mm. of mercury, probably beoau•e the lellk b decreased elightly to maintain constant total diachaf"ge-ourrent. s11oh variations are second order etfeot1, howeffr, and in general the number of' atn.y eleotrons 1e l)l"oportional to the beam-outtent if preaeure ooncHtions within the tube are not n.rie4. The diatl"ib~ion of atn.y eleotrone in. the -various parts ot the oaoi1lograph ftried at dittel"ent timea, ud th!.• varlati011 could not a.l•ys be correlated w1 th ftrla.t1ons of ga1 preaeure at ditteNnt part• ot the tube. Appa1"ent1y tlua diatrl'bution of •tray eleotl'oits 1e alto attected by the intensity of the oul'Jl'ent in the beam, but no detinJ.:te rel.atlon wu determinect. 53 40 ~- £1) w ~ 4:. ' . ' - --- - · - · · · · - - - -- . ·· ·- - ·- · · · - - ·· · · · · ·· - · · - -- -· - . . ... l ! _ __ _ .... FIGURE 22 - - - AT AGE 3 6: 09 ·•- t•• [ - B - S TRAY [L[ CT RON CURRENT ADDED TO TRAP A - Bf AM C URRENT LIQUID AIR BEAM CURRENT. EFFECT OF WATER VAPOR ON · ·· • ·-· -+-- · - - -- · ·-·-·•- ·· - • 36 37 CATHODE AGE IN 38 HOURS 39 40 Q..___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _...,___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _- - - J 10 •· - - 3 20 a:: er t2 w I :a 30 c..:> a: 0 <C ~ u.1 Cl. w a: U) 50 60r-------~-------r--------"T-------, . -- - - -- --r--- I 0 )0 ! w co - _ j _______ _ _ i i I ~20 <( u :::::i Ct'. w 3 0 t - - - + - - - - - - - - + - - t -· -- z er: ..... ~ u <t. ~ ~401' a.. 2 w o::: f3501-+--r-- 60t------ct-r--··•·-- -- --- . . - . ~ , AGE 4 I ' i FIELD DURtN(; RE.i\Oll\lGS .. I -·· ·-- --· ··---···- -·--··- ---·· -·-··-·--- -·· - C -NO FORE C.0NCfNTRAT1GN B - FORE-C0NC£NTRATI0N FIH D ALMOST CONTl'NUOlJSLY AT ANY TIME CATHODE A -NO FORE-CONCENTRATION FIELD - FIGURE 2.3 - - -· L._.. BEAM CURRENT - - + - - · - - - - - -- 2 3 CATHODE AGE IN HOURS · - ·- - -- -- -• . .r -- i r-·-- 5 ?Or---------,---------,----------i--------,---------~ An auxilliary leak valve in the lead between the osoillograph and the vapor pumps allowed tests to be made at pressures higher than the normal operating pressure of 2 x 10-4 mm. of mercury wl.th the needle valve leak in the discharge tube the major sourGe of gas in the tubee The first visible change in the appearance ot 1me beam ooe\U'red at a p:reesure or 5 x 10•3:mma stray electron illumination or the baokground. in~reased. notieeably and the edge of the beam became lese 1he.rply detined. At 1 x 10•2 mm. air pressure the tube was still operative, but transient de- focusing •s quite bad., and the sweep voltage tla.shed over to e;rouad frequently. This behavior -was identical to the opei-at1on of the tube with water vapor present at about the 1111ne pressure (1 x 10•2 mm.) and a muoh lower partial pressure of the air itself. Cathode surtaoe conditions are a second taotol' which influenoes beam efficiency. A cathode 'Which has not been in operoation tor some time gl"adually regains its efficiency if operated oontinuously. The loss ot efficiency depends on the length of time the eathode has been inoperative. The disoontinuitiee in curve A, Figure 23. represent efficiency changes caused by removing the aooelerating voltage fol' short intervale to de• tennine whethe:r the cyolio variations or etfioienoy wel'e oaused by con• tinuous Operation. The break at .time 15 minutes was not longe~ than 5 minutes in duration, while the brea.ka at 2 hours and 4 hours lasted approximately 30 minutes ea.oh. The water vapor trt;p was filled with liquid air continuously during these tests. on several other- occasions, after the tube had been inoperative ovei- night, voltage was applied intermittently until no further inorease or bea.m-.current or deorease or st!"ay ~leetron~ current was obeerved, to aesure that the tube was free or water vapor. Then voltage was applied oont1nuoua1y, and the beam-current and stray 56 electron-current both increased during an interval ot one hour. The beam-current nearly doubled during the inteM'e.1, and then heoa.m& noad7 at the value of beam-current obaerved during the tests made on the pN-rious day. This behavior varied with different cathodes, but does indioate a relation between bea..~ etficienoy and cathode use. The cyolie variations or beam effioieney -which ooou:r duf'1ng the first few houl"'s o:f' cathode age a.re illustrated in FiguN as 'by the oul'Tea or beam-current as a :f'unction ot cathode age, with total disoharg•-ouM"ent maintained oonstant at 3.0 milliamperes. curve A gives the beam...o\U'l'"ent from a cathode which was aged without using the fore#oonoentration field at any time. Data tor ourve B were obtained by using the fore-oonoentre.tlon field current .~,hioh ga~e ma.xiraum beam increase at all times except whil• data for curve C were being obtained. The needle va.lV'e leak was re• adjusted eaoh time a change was made trom. curve B to ourve C so. that di•• charge-current was oonsta.nt a.tall times. No wide nriations of needle valve setting; 'W8N made 'While obtaining the data. fop eurft A. Earlier tests indioated that there might be a. systematic increase ot total clis- charge-euttent when the beam-current 'WS.S inoNaaing, and a less definite decrease of total discharge-current when the curve had a negative slope. This was corrected by adjusting the needle valve leak slightly to main• ta:tn a. constant total disoharge•outtent. Attempts 't(? cheek suoh a TU'ia:tiMl of the total disoharge-o'IU6Nnt weM unaucoe&$ful 'because the nMdle Talv•. assembly developed an erra.tio leak whioh masked the smaller Tal"iat1ont ot total discharge.current. Allowing the cathode to age without ow:r-eoting the changing disehe.rge-eurrent does not elimina.te the cyolio efficiency vanationt. If the oathode is moved, even slightly. after the beam efficiency has beoomo steady, the beam ..ourrcnt will decrease greatly, then recover as a 57 new erate:r is formed by eroding a. side of the old crater'. lfo evid&no• o:f more than a single f'luo'tua.tion was observed., although it ia po•rdb1e that several might oocur. A crater does not form immediately when a new cathode is in- stalled, but forms after a delay of approximately 6 minutes. Thie time corresponds to the firet sharp maximum. in the beam... o\U"rent O'Urve. independent experiments oentirm this obtervation~ Two The surtaoe of the cathode can be observed, using the oscillograph as an electron miorosoope, and no indication of a crater exists dul"ing the :first rew minutes of use. Or., the oathode may be removed af'ter a fff m.in,:rtes use, attd a matt• surface has been formed on the cathode by the bombardment or the poaitiw ions, but a crater has not yet been developed.. The beginning of a crater is also dependent on the condition ot the cathode sur-faoe when a disoharge i8 tirst started. One cathode gave a beam cn1rI"ent or 38 mioroamperes at an age ot 1 minute, a ma:itURlm of 62 mioroampe?-ea at an age o!' 5 minutes, 34 mierosmper-es at an age ot 10 minutes, and 8 mioroamperea at an age of 16 minutes. Another oathode, which was a.ooidentally subjected to a high-.preasure glow discharge wlum first installed, gave a. blarll•OU?Tent of 8 nt.ioroamperes at an age ot 2 minutes, :reached a maximum. of 30 mioroa.mpe:Pes at an age of 13 m~iJ, and the beam-eurrent had decreased to 24 mict-oamperes at an age of 18 maut••• Both oathodes viere aged wlth a to-bal dis.eharge•o'I.U'"rent of 3 .. 0 milliampel"es. Both oathod&s gave a stoady beron--ourrent of 26 mio!"OampeNls after th-, had been operated continuously tor 6 hours. These data refer to the beam• ourrent without any fore-oonoentration field, although a fO~e-conoentration field was used at other times dur1%lg the teats. The weight or material lost by the cathode as a Ol'ater is formed. is proportional to the time the cathode has been used, and within the 58 possible aoeura.cy of m,aaureme nt, seems to be the _same for a cathode aged. without fore ... ooncentration, a.nd a oathod.e aged with a fer--e-cone.e ntr-ation field applied continuously. Average weight o:f' the cathodes u.aed ns 170 milligrams ~d the wight loss was about 1.5 milligrams d'Ul"ing 10 hooios of operation. The sputtered materi al is deposited on the oooled -.11& of the dis charge ohamber, probably in t he f'o:rm of alum.inum <»tide ., W'R I TI:HG SPEED Three taotors may l i mit the usef'ul writing speed ot a oathod.e ....N.y osoillograph. The eleotron energy may be too low to permit ss.tisfa.etory observation or photography with the beam. The transit time of the eleo't!"O?lli past the d.etleetion system may become long relative to the period ot the voltage vai-iation being measured and introduce .&.Jlll)litude distort.!@. It a sinusoidal voltage va.l"iation of a period equal to the tr-anait time is applied to the deflection plates, this distortion will produoe iero de- flection. The third faotor ia the natural trequeney ot the detleotion system i tsalf'. Any instru.rn.ent is limited a.$ e. measuring device it the period of the l:!1$asured variat i on oorn>aponds to the natural period of' the i nstt'Ument. Those three factors 1:>ecome important almost eimultaneously in the oscillogra.ph under consideration. The electl"on energy avaHable pemi:ts a writing sp♦ed of 2 x 109 0111. per second .. This writing epeed is obtained if the recorded amplitude of osoillationa at the natural trequenoy ot the deflection plates ud leads just tills the f'1eld of the cam.era~ The transit time of the eleotront!f, neglecting relativhtlo eorfiections wh:lch a.re not great for a bea.'ll with an energy ot 10000 eleotr-on volts, oan be obtained rrom, 59 im,T2 • eV • T • a!.v • 6 x 109 am/seo. m. wheres is the length of the deflection plates, in this case, 6 •• The natural trequenoy of' th.a defleotion synem 18 a.pproxtmately 125 megacycles per second, so the natural per'iod iss T • ½ • . •. 1 .. ! • 8 x 10•9 aeconcit 125 X 10 and the transit time ot the eleotrone in the beam is sufficiently l!lmt.11 to prevent amplitude distortion tor any trequenoies which can be meae.urecl with the osoillograph. The maximum wl"i ting speed was obtained by oonnecting both de• tleetion plate terminals to a emall capacity with the shortest possible leads. Coupling leads to the blocking gap, 1so1at9d by coupling condensers, were connected directly to the detleotion plate terminals to obtata Figun 24, and m,re ooup114 to the deflection pla:be terminals by small ca.paoittes made by twisting t'WC insulated wires together when Figure 25 was photogftphed. Disoha:rgt•ourrent for- the1e photogl'B.phe wu s.o milliamperes, and the cur• l"ent in the beam, with f'ore-oonomration, was 88 mioroampe,tes. A sweep circuit ten <times taster than the fastest sweep noi"mally e:n:1.lable with the oscillograph was constructed tor these tests. Writing speed is def'ined a.a the n.tio ot the length of the trace to the coneapOJtding value ot the elapsed. time. Foi- a sinusoidal wave form, the muim:um m. ting speed may be computed f'l-oms i •k sin A l,rft • 2ntA cos 2n:t"t wh•:re A ie the amplitude et the detleoticm.. The above equation givee the correot maximum writing speed only if' the tN.ee is ntuly perpendicular' to 60 Figure 24 125 megacycles Natural frequency of deflection plates Figure 25 125 megacycles Natural frequency of deflection plates the direction of sweep. Substitutlcn or 125 mege.eyclea per second tor the f'requenoy gives a maximum 'Wl"'iting speed somewhat greater than 2 x 109 am./eeo. tor Figuzi,e 24;, and 8 x 108 om/aec. fo'1! Figure 2&. values are greater than those quoted by R. H. Grleat1 . is due primarily to the l"'eduotion or beam diatort1.o•• 62 Th••• The improvement RECOMMENDATIONS One eonclusio:n is obTious - a better vacuum system is neoeas•J7'• The present mercury vapor pumps might be imprOTed by cleaning, and would be improved ii' je'ba of newer design were installed. But the a.dTan-b.ge.s or modern oil ditfudon pump• make the replacement or the pres<Jnt pumpa desirable. One precaution is necessary. 011 ditfud&n pumpe aN damaged. and may even become ittopel"t\tive if a luge amount of' water vapor comes in contact with the heated oil. cenain oil diffusion pumps are selt- f'l"aotiou.ting and might be used to Nmove the moisture .from the til.ms if' the oil ,.,,el"e Nplaced .floequently. Using the present vaouum pum.pa to N• move the water vapol" trODt the films before they a.re plaoed i:n the oscillograph, and 1nata.11at1on of a new, lU"gei- tore.pump and ad.equate oil diffusion pumps would be preferable. A liquid air vapor tnp at the ail" inlet to the nef>dle valve leak •uld also be desil"abla,. as 'the humidity of the atmosphere beeomea quite high oooaaionally. The cathode assembly could be improved by subst1-tuting a glass to kovar seal for the present mioa.-picein ••al. Thia change might be aO• compliahed without any :revision of the metal parts of the present assembly. Elimination of 01te aouroe of Taouum trouble and more exact and permanem centering or the cathode would be the i-eeult. Substitution of rubber gaskets tor other wax joints should be considered. Additional data on the behavior of the discharge as a cathode ie aged are desirable. The aluminum deposit on the walls ot the discharge tube has been Nmoved ainoe the data for FiguN 23 wre obtained~ Similar tests of be-.m-ourrent as a funetion. of cathode age should be. repeated in order to dete'tmine the effect or the ohanged dimensions of the disoharge oh.ember 011 the anomalous behavior 0£ the beam-ournnt.. 63 rt the phenoiaen.a pet!'sists, a. series of oathodes might be removed at ditf'et'ent age• and. ex• amined miaroeoopioally to determine whether the shape or the oath.ode crateivariea eyclioly with the -re.riations et bea~ et.fioienay. !nfo1"mat1on re• garding the behavior· of a cold cathode discharge mi~ht be obtained troa continue,j investigation of this an.oma.ly. 64 BIBLIOGRAPHY 1. R. H. Grieats A Low Voltage High-Speed Cathode-Ray Osetllograph. Thesia to., Ph.D. degi-ee, CalifoJ!'nS.a. Institute of Teehnology., 1937. 2. P. F. Hawley, Design and Opel'at!.on of a Cathode-Ray Oseill-,graph tor External Photogra.phy. Thesis tor Ph.D. degree., Califoi"lda Institute of Teohnology, l9S6. 3~ J. G. }>1easants: Observations on the Theory and Charaot♦rtst!c• of' Electrical Figures en Plates in Ai~. Thesis for Ph.D. deg:ree. California Institute or Teohnology, 193$. 4. w. G. Dow: FUndamenta18 of' Engi:n.eerbig Electronioa.. and Sona, Ino., New York, 193?. 5. w. R. Smyth$: John Wil•y static and Dynamic Eleotr1c1tyd McGraw~1u11 BQGk Company, Inc., New Yo2"k, 1939. 6. L,, B. Loeln Fundamental Prooeasee or Electrieal Disoharg•• in Gaees. John Wiley and Sons, Ine., New Yo~k, 1919. 7. L. M. Myens Eletttr>on Optics - Theo.retieal and Praotioal. Chapman and Hall; Ltd., London, 19$9. e. I. G. Malott and D. w. Epst•m, ElectP011 Opties in Televisioa. MoGre.w-Hill Book COIIlpflll¥, :mo., New York, 1938. ~?. A. B. Wood: Cathode.Ray Oeoillogr,aph. P• 1046, Novembep l9SS. I. E. E. Jou1!'nal. Vol. $S, • 10. J.B. Johnson, The CathQd.e•Ra.yOsoill«>gra.ph. Joumal or the Franld1n Inatitne, Vol .. tli.t P•P• 687•?17, Deoember 1931. 11, M. von Ardenne1 Untersuchen an Bn.UJlSohen Rihren mit Gasrilllung. Hooh.frequencteoknik und Elektroakustie., Vol. 39.t pp. 18•2•, 1932. 12. II. Busch, 'Obet" die Wir-kungsni•del" Konsentri•rungsepule bet der Braunsehen R&hlte. Arohiv f'6:t" Elekbroteehn!k., Vol. 18 1 P• 583, 1927. 13. v. K. ZWOl"yldns 14. E. Ruan, Bau~~ ;~~stung 4•• magnetisohen Elektronenmila-oskop•. Zeitachritt tur Physik, Vol. 87, PP• 580-602• 1934• On Eleotron Optics. Vol. 215, p. 535, May 1933., 66 Journal or the Frankl.in Institute, .APPENDIX A "Ion1&ation Times of Thyratrons" By Arthur E. Harrieoa AIEE Technical Paper 39-116 Apri I 1939 IONIZATION TIME OF THYRATRONS ) A. E. Harrison Enrolled Student California lns~itute of Technology, Pasadena Advance Copy Not Re I eased for Pub I i cation A I I R i g ht s Res e rv e d by t h e American Institute of Electrical Engineers, 33 West 39th Street, New York, N.Y. A paper re~omrnended by the AIEE subcommittee on electronics and scheduled for presentation at the AIEE combined summer and Pacific Coast convention, San Francisco, Calif., June 26-30, 1939. Manuscript submitted February 16, 1939; made available for preprinting April 25, 1939, IONIZATION TIM!!: OF THYRATRONS By A. E. Harri son SUMMA.RY enteen thyratron and grid glow tubes, representing 12 different types, have been and their ionization times have been measured with a cold cathode type oscillograph. nizat1on times vary over a wide range, from a fraction of a microsecond to several inicroseconds. Ionization time has been found empirically to be inversely proporo grid overvoltage. Grid overvoltage is defined as the difference between tha l applied to the grid and the critical potential necessary to p1~vant firing. Argon, d mercury vapor tubas were included in the test. Mercury temperature was found to appreciable effe~t on the ionization time with low grid overvoltages. All mercury bes have similar ionization time characteristics. In fact, tubes of the same type facture sometimes differed from each other more than one tube dif~ered from anothe r different construction and manufacture. d controlled vapor filled rectifiers, or thyratrons, have not been considered in t he use in blocking and sweep circuits for cold cathode type oscillographs because of herent ionization time . A search of the literature on the subject, and reference t o urers' data sheets furnished little useful in:f'ormation on this ionization time . A program to determine the minimum. ionization times of various tubes was thus start ed and first results showed that minimum. ionization times of 0.3 microseconds were le with mercury vapor thyratrons if the grid was driven sufficiently positive. These suggested that these tubes might be successfully used in blocking circuits for high cillographs, and work on this phase of the problem is being continued. The apparatus for measuring ionization times made possible an extended study of the relation benization time and grid overvoltage a s a function of anode voltage and vapor temperature, cury tubes) and it is the results of this extension of the original problem which discussed. EXPERIMENTAL PROCEDURE .ization time of a thyratron, or the delay between the time of application of a square ulse to the grid and the time the tube voltage reaches its normal conducting value, y measured with a high speed cathode ray oseillograph arranged to record the voltage he tube, provided that the time of application of the grid impulse is also shown. cessary to synchronize the unblocking of the high speed oscillograph with the tripulse applied to the tube under test . (Tube Bin Figure 1. ) This is accomplished .s e of a trip tube (Tube A) which is fired by a surge from the oseillograph blocking delayed a fraction of a microsecond by R5CG• An impulse of magnitude determined •atio of R3/ (R3+}¼) is applied to the grid of the tube under test when tube A fires. ulse voltages from 3 to 63 volte are selected with a tap switch connected to differes of R3 , allowing the grid overvoltage, or difference between the potential ap, the grid and the critical potential necessary to prevent firing , to be varied over •ange. The time of appl ica tion of grid impulse is recorded on the oseillogram through ,ling capacity c4 • C4 and C5 are small coupling capacities fo:rmed by l inch of twisted insulated wires • .ues of grid impulse voltage were calculated from the ratios of R3/(R3+~) for the taps and were checked with the cathode ray oscillograph. The apparatus was mounted the oscillograph as possible (See Figure 2) in order to reduce the length of le.ads • .th the oscillograph verified that the time constant of R4 and the grid to cathode , of tuba B had no appreciable effect on the measurement of the i onization time. ) -2First tests were made with the tube under test in a circuit similar to that of tube A without R5 , and ionization times were measured from the beginning of the oscillogram. Two factors made such a simple circuit unsatisfactory; there was no control of grid overvoltage independent of initial grid bias, and tests with an 885 indicated that the delay in the blocking circuit of the oscillograph was greater than the mi.nimmn ionization time of this tube. Both drawbacks were eliminated by the addition of a trip tube to the circuit. An 885 tube was finally chosen as a trip tube because its breakdown time was quite short and gave the best approach to a true square wave impulse. B5 was therefore added to delay the trip until the oscillograph sweep had started. Direct coupling and capacitive coupling to the grid of tube B were tried and capacitive coupling was finally adopted because it allowed control of in.1tial grid bias with a potentiometer and frequent check of the cri ti-cal bias was possible. Originally, carbon resistors were used for R3 and R4; but this type of resistance has an odd behavior under transient conditions and the grid impulse voltage could not be calculated from the d.c. resistance of the carbon units. Metallized resistor units for R3 and ten metallized resistor units approximately 400 ohms each in series for . R4 overcame this difficulty, giving a voltage divider with approximately the same characteristics for d.c. and transient conditions. Condensed mercury temperatures were measured with a mercury thermometer wrapped with tinfoil and tied against the base of the bulb. This method did not give accurate measuremant of the temperature, and did not allow tests at subnormal vapor temperatures, but the results do indicate qualitatively the effects of changes in mercury temperature. Test data were obtained photographically for the 12 types of tubes listed in Table I after the tubes had reached an equilibrium temperature. Three or more oscillograms ware taken for each value · of grid overvol tage and a fixed anode "\101 tage. The sensi ti vi ty of the oscillograph was such that anode voltages greater than 1000 volts could not be recorded easily. The time base was calibrated with a radio frequency oscillator which was checked against broadcast stations. Figures 3 and 4 are representative oscillograms of anode to cathode potential. Ionization time is measured from the time of application of grid impulse, recorded as a slight rise in voltage at the beginning of the oscillogram, to the time at which the tube voltage has dropped to its conducting value. The ionization time for th& strobotron was an exception, since a strobotron is used without a load resistance to discharge a condenser, and its ionization time was measured to the beginning of breakdown. EXPERIMENTAL RESULTS Ionization t:l,.me as a function of grid overvoltage has been plotted in Figure 5 for six of the tubes listed in Table I. Inspection of' the curves of Figure 5 suggests a resemblance to a family of rectangular hyperbolas. Figure 6 was prepared to check this suggestion and shows the reciprocal of ionization time plotted as a function of grid o·vervol tage. The resulting curves are nearly linear, indicating that .the hyperbolic f'orm of the curves in Figure 5 is not an illusion. The greatest deviation from linearity is shown by the FG-81; but this is not surprising since the maximum grid overvoltage is high compared with the anode voltage. All of the argon filled tubes reached a low value of ionization time with a grid overvoltage of 20 or 30 volts, and further increase of grid impulse voltage failed to reduce the ionization time appreciably. V. C. Rideout, graduate student at the California Institute who is continuing this stud~ has varied grid bias, grid impulse voltage, and anode voltage proportionately in an attempt to maintain proportional gradient values for each test. His conclusions are not yet definite, but there is reason to believe that this technique gives reciprocal c.urves which are more nearly linear~ Published data regarding the ionization times of thyratrons are meager. Hans Klemperer(l) in 1932 suggested a theory based on the mobility of the positive ions, and his calcula·t ed icnization times agreed in magnitude with his observations. Unfortunately the magnitude of the positive impulse applied to the grid by his method was not given, although it apparently was large. Xlemperer' s theory predicts that the time from the application of grid overvoltage until the start of voltage decrease across the tub&, -{Ti in Figure 4) and the time for the anode voltage to reach its final conducting value after decrease across th& -3- tube has begun, T2 , should be approximately equal. This relation is true for large grid overvoltages, but it is definitely not true for low grid overvoltages. The ionization time may depend on the inittai ~~ber of electrons available to cause ionization, the so-called dark electron current.C 4 ),l 6 J Some vapor tubes have much larger dark electron currents just before breakdown than others. Figure 4 indicates the existence of a relatively large electron current by a gradual drop in anode voltage before final breakdown. No change in the ionization time of a KY-21 was observed when R2 was changed from 20,000 ohms to 2,000 ohms if the grid overvoltage was greater than 5 volts. There is a substantial reduction in the ionization time for grid overvoltages of 2 or 3 volts, however, when the load resistance is reduced. As yet, a theory to predict or explain the observed relation between ionization time and grid overvoltage has not been suggested. Observations regarding individual tubes may be interesting. The electrode construction of an 885 differs radically from that of the other tubes and this thyratron has a very short ionization time, practically independent of anode voltage. In comparison with an 885, a 2A.4G has much longer ionization times at low grid overvoltages, but the ionization times are almost identical at the higher grid overvoltages. Compare Figure 7 and curve A in . Figt1.re 5. An FG-81, however, has a ionization time curve more nearly like that of the mer- · cury vapor tubes, although it is argon filled. Ionization times for neon filled tubes were in all cases longer than the ionization times of argon or mercury vapor tubes. Two KU-610 tubes were tested qd gave slightly different results with 750 volts applied to the anode. These tubes would not operate consistently at anode voltages less than 500. These were old tubes, however, and this behavior may not be caused by the use of neon gas. Direct comparison of tubes filled with noble gases could not be made because the ionization times are afi'ected by gas densities in the tubes and these were not known. ill tubes except the 885 show a variation of ionization time characteristic with different anode voltages. In all cases where anode voltage affected the ionization time curve, an increase of anode voltage caused the knee of the curve to be more pronounced. Figures 7, 8, and 9 illustrate this effect for three types of tubes. The longer ionization time for large grid overvoltages with 100 volts on the anode is due primarily to an increase in the breakdown time (T2 in Figure 4), as the time lag Ti is practically the same for all anode voltages when grid overvoltage is large. Note that there is little difference between the curves for an FG-5'7 and a KY-21 in Figures 8 and 9. A second FG-57 showed ionization times consistently longer than those for the KY-21. It is common for mercury vapor tubes of the same type and manufacture to differ from each other in characteristics, more than one differs from another tube of different construction and manufacture. No change in the ionization time of several mercury vapor tubes was found when the tube was biased to a large negative value (-50 volts bias for a tube with a critical bias of -5 volts) and the positive grid impulse increased so that the grid overvoltage. was the san:e as before. Other circuits may give ionization times much longer than the values given here if they use high grid limiting resistors which introduce an additional RC lag. An unusual phenomenon was observed while testing an 866 which had a special shield electro~e placed around the cathode. The critical bias necessary with 400 volts on the anode changed within a few hours use from. -4 volts to -60 volts. Loss of grid control continued as long as the filament was heated, although the anode voltage was zero and the grid bias -60 volts. A. w. Hunl 2 J mentioned this effect in 1929. The tube mentioned above was new, but the construction may be incorrect for use as a thyratron. In some tubes, particularly the FG-17, a variation of ionization time for a given low grid overvoltage and low anode voltage was observed. The special 866 and the KY-21 also showed this effect, but the spread of ionization times was not as wide. An anode voltage of 300 reduced the variation, and 1000 volts on the anode practically eliminated the effect. ill of these tubes have filament type cathodes and the effect is probably related to changes in effectiva cathode potential for different parts of the cycle. A difference in emissivity for different parts of the filament would exaggerate the effect, although the center of the filament supply might be at ground potential. The FG-17 tubes were several years old and probably not in the best condi tion, which may account for their pronounced variation of ionization tiroo for a given low grid overvoltage. ) Variation of the ionization time characteristie of an FG-57 operated witn an anode voltage ot 300 is shown for three different mercury temperatures in Figure 10. Mercury temperature has considerable effect at the higher grid overvoltages, but the curves coincide for low grid overvoltages. The temperatures indicated are not necessarily accurate, as they were measured by a questionable method. In fact, a series of tests with different temperatures and a fixed value of grid. overvoltage gave erratic results. But each of the three curves in Figure 10 represents a consistent set of data for some temperature, and the trend is certainly definite. Changes in temperature produce a similar effect on the ionization time curves of all the mercury vapor tubes. Two cold cathode tubes were tested. The OA4G is argon filled and its maximum ionization time is very long. The strobotron is a neon filled, cold cathode tube, and cannot be compared directly with the other hot cathode types. It is interesting to note, however, that the curve for the strobotron resembles the curve for the KU-610, a hot cathode neon filled tube, if the curves are displaced vertically until the minimum ionization times are equal. (Figure ·5). Perhaps this is an indication that there is a fixed time lag necessary for initiation of emission, although the data are not sufficiently complete to prove this theory. Two values of capacitance were used across the anode and cathode of the strobotron and the load resistor (R2 in Figure 1) was zero. Two microseconds were required to discharge a 1 microfarad condenser, and eight microseconds were required to discharge a 10 microfarad condenser. These values are the time required for the condenser to discharge. Discharge occurs after a delay, depending on grid overvoltage, as given in curve F, Figure 5. ) The results of these investigations show that the ionization time of mercury vapor thyratrons may be as short as 0.3 microseconds if a positive impulse of sufficient magnitude is applied to the grid. Certain argon filled types have even shorter ionization times. Characteristics of tubes of the same type may vary more than characteristics of'tubes of slightly different construction. All tubes show the same general relation between ionization time and grid overvoltage. For most tubes the ionization time is inversely proportional to the grid overvol tage. There has been no ·theory suggested to explain this observed relation. However, these empirical data can be the basis for practical application and further investigation. BIBLIOGRAPHY 1. UBER DEN ZEITLICHEN VE.RL.A.UF DER zfurrnJNG VON Io:NENROHREN, Hans Klemperer, A.rchiv ftlr Eleotrotechnik, volume 27, 1933, pp. 322-328. 2. THYRATRONS, HOT CA.THODE, A. W. Eull, General Electric Review, volume 32, April 1929, pp. 213-223. 3. IONIZATIO.N Til..W. OF THYRATRONS, L.B. Snoddy, Physics, voll.lll18 4, October 1933, pp. 366-371. 4. P~-STRih.'ING ~RENTS IN A THYRA..i."'RON, E. T. Wheatcroft, R. B. Smith, J. Metcalfe, Philosophical Iv1agazine, volume 25, .April 1938, pp. 649-663; also z. T. f;neatcroft and T. G. Eammerton, Philosophical Magazine, volume 26, November 1938, pp. 684-694. 5. ~~~:RONICS AED ELECTRON TUB!l:S, a book, E. D. McArthur, John Wiley and Sons, Inc., 6. :F'Ul\JDAI.~JTALS OF El-!GTh"EERING EL:i!:CTROHICS, a book, William G. Dow, John Wiley and Sons, Inc., 1937. Includes a good bibliography. 7. ADDITION;lL BIBLIOGRAPHY ON V,U'OR 'IUB!l:S, General Electric Review, volume 41, October 1938, page 456. TABLE I RANGE OF IONIZATION TIMES - MICROSECONDS Number of Tubes Tested Tube Manu.f'acturer Gas .Anode Voltage Shortest Measured Ionization Time Approximate Maximum Ionization Time HOT CATHODE TYPES 2 FG-17 General Electric Mercury 1000 0.50 50 2 FG-57 General Electric Mercury 1000 0.40 50 1 FG-6'1 General Electric Mercury 1000 0.60 150 287-A. Western Electric Mercury 1000 0.65 50 l KY-21 Eitel-McCullough Mercury 1000 0.45 20 l 866 Special * Mercury 1000 0.35 20 3 885 RCA. Argon 300 0.08 1 l 2A.4G Raytheon Argon 300 0.10 10 l FG-81 General Electric Argon 180 0.25 50 2 KU-610 Westinghouse Neon 750 l.80 300 l VlE COLD CA'IlIODE TYPES l OA.4G 1 Strobotron Raytheon Argon :300 3.0 300-1000 General Radio Neon 350 12.0 300 * An 866 r&ctifier with a control electrode added. O·IOOOV OSCILLOGRAPH TRIP - 250 V Manufacturer unknown. Figure l. Circuit diagram of thyratron test apparatus R1 = 1,000,000 ohms R2 = 20,000 ohms R3 = 50-1350 ohms R4 ~ 3,?00 ohms R5 = 100,000 ohms R5 = 1,000,000 Olli~ B R7 = 1,000,000 ohms Rs• 1,000,000 ohms C1 = l mtd. C2 = l mtd. C3 "' 0,001 mfd. C4} Twisted wire C5 capacities CG= Electrode capacity Tube A - 885 Thyratron Tube B - Tube under test ) 0 .0 0.2 0.4 0 .Co Q.8 MICRO.SECONDS l.O Figure 3. Typical oscillogram for an 885 thyratron T,-,Tr T--------~I Figure 4. Different phases i n breakdown of an FG-5? T1 Time l ag T2 Breakdown till8 T c Ionization time Figure 2. Thyratron test apparatus connected to the hi gh speed oscillograph i I\ 1--- .,..v \ 15 ) 3 \ F I"' ....__ <f) Cl z / r-- c--- \ 0 0 w /v w ::, B/ V A i== ·Z <f) 2 2 ~10 ~ 0 I/ ,v I I/ N ::, z I\ w \ ::, i== \ \ z \ 0 ~5 N z I 0 \ ' \ \ B I'-. 0 A - V ..J <( 0 E a:: a.. I (.) w I a:: \ ~ .........._ I\ "' "' /v "-... C/ / I r--... / / >--- " I_____.,. 40 GRID 0 V ER V0LTAGE / I 60 Figure 5. Typical ionization tiim curves ~or thyratrons A - 885 Argon filled thyratron 300 volts on anode B.- FG 81 Argon filled thyra tron 180 volts on anode C - KY 21 I.ilercuxy vapor thyratron 1000 volts on anode D - FG 67 Mercury vapor thyratron 1000 volts on anode E - KU 610 Keon filled thyratron ?50 volts on anode F - Strobotron - Neon filled, cold cathode 350 volts on anode 0 V / V 1,/ V / V E.....- V V ./ - // --- -- o/ / / V // I / ......__ I'--- r--. 20 / (.) \ !~ V I 0 ___,..... ----- ,__ i---- i--20 GRID 'IJ E0 0VERV0 LT AGE Figure 6. Reciprocal ioni zation times for Fi gure 5 A - 885 Argon filled thyratron 300 volts on anode B - FG 81 Argon filled thyratron 180 volts on anode C - KY 21 N'ie rcuxy vapor thyratron 1000 vol t s on anode D - FG 67 ::Iercury vapor thyratron 1000 volts on anode E - KU 610 Neon filled thyratron ?50 vol ts on anode 15 15 Cf) Cf) 0 0 z z 0 0 w w 0 0 Cf) Cf) ~10 ~10 0 0 ::!: ::!: w w :IE j::: ::!: z z \ \ ~ 5 ~5 \ z z j::: 0 0 ~ N ,\ '~ K.. 7 I\ \I\ \ 0 0 \ I\. 100 VOi TS 300 VOi TS ,.__ 20 40 GRID OVERVOLTAGE 0 60 ". '~ - 100 VO TS l.:>UL vu T S '\. ~ boo VO TS ~ I'---.... 20 40 GRID OVERVOLTAGE 0 60 Figure 7. Effect of anode voltage on ionization tiw~ of a 2A4 G argon filled thyratron Figures. Effect of anode voltage on ionization time of an FG-57 mercury vapor thyratron mercury temperature 70° C 15 15 Cf) Cf) 0 0 0 0 z z 0 0 w w Cf) Cf) ~10 ~10 0 '-' I ::!: ~ ::E I w 1- z 0 ~5 !::! z Q ::E j::: \ \ ' I\ z 0 ~N 5 I\ ~ k, hoo \ \ '\ 1"'0 ~ z "- ~ i---.. ..._ r-~ - 20 40 GRID OVERVOLTAGE \\ \\ I\ 'l'\\, Q 10 0 VO TS !300 VO TS VOi TS .. ,\ .\ w ::!: . - 50° C 70° C ~ I\.) "-L..----- so ·c ~ 1'." I'--.. 60 Fi gure 9. Effect of anode voltage on ionization time of a KY-21 me rcury vapor thyratron mercury temperature 67° C 0 ~o 20 GRID OVERVOLTAGE t0 Fi gure 10. Effect of condensed me rcury temperature on ionization time characteristic of an FG-57 with 300 volts on anode ADDITIONAL DATA ON 11 Ionilation Tim.es of Thyfttl"ons" I I I - I •--·· ·- ,- ---·· I. . . . ... . . . ! I I t· · I ! ... - . --- I ti) I 0 2 . i - -~ --------I 0 I I I -- - - - -I -· - -- - +-- -~- ~ \0 <./) 0 i er u I l I 1 II ! i . ,- -- ----- - - t - -- .l . - - 1 • ~ ' w ~ 1- z 0 ~ 5 -z N ! - + - - - - - - ~-- - - - - + - - - - - - I - I 0 i 1oo !voL s -- ---- -- - - --+---- --·-· -- ·•- 300 i VOLS I _ 100_9. f_'{Q_h __ ~ --- 0 20 40 60 GRID OVERVOLTAGE IONIZAT\ON Tl MES FG17 54°C -· 15 - · ·- · I I I I Ii I l I ·- I ~ - I I 1 if) I I 0 i 0 I z i u w 10 .1 - - - .... <;/) 0 ~ ·- (..) 2 \ w \ - ·- ,.,.- - ··- ·-- <( N 5 -;z. \ I I I --- - -- - \t -·--· I -- -----·---,-. ·---j --t- I\ -·-- :;- I ! - - -·- ,.__ i I --H ! I i ' I i i I I .t-- - ·- - ~ -·- r . - - --·-··-- --·- - ---~--+ I \ \~-\ --~ \ ~ - -/ ·-------' t¥ -·. "' - - -· - ------- - 0 I I I --·-- ·- - ' 0 ~·--+--i I - 0 I I -r-I : ~ z I I 2 I- - I I I 1S--~ ..,-~- ,oo~VQ_L ---~ r-,..... ~ 1000 / ~~ · - ··· r------ - - - - - 300 VOL TS - ·· - - r - · · ·· ·- • VQ_h T.S - - ---- ~ - - --·- " I'----~r--- -- ·---~ -- - - . t .~ . ~ ~- -t:--,,. -· - -- --· --- - 20 40 GRID OVERVOLTAGE IONIZATION TIMES WE 287A .. ·- 60 55°C I i 15 1 - --+---t- - - - - - + - - - + ---i--i- - -- ~ - - r - · - - + - - + - -- f - - - - - - + - - - - + - - ~ --- ---·-- --··-- ----- -~- -- --- - ....... ! 1----·--· - 1 - -- - - -- - -- - - ····- · ··- ·- -- -- - -- - I i' -- - ... .. I i I : ~+ - --- -- c/) 0 - -l- - - - I - ---- -:- -- -- -- __·--____. _ - --- -- -- - ----+---+--~ :z 0 u w 10 - + - --- - - H - - - + - - - - -- - - - - 1 - - - - - - - -- - -- -- --+--------1 <./) 0 ~ u ~ - ~-- -- t-- --·- i w ~ f - + + - - - - t - -·- \--l ._ _ - - - -~- l_______ ~ i - t - ---4 - - - -- -- -- + - - - - - - - - + - - - - - ! i -- - - - - 4 - - - - -- - · - - - - -- - - - ·-+-----l t- z ·- - - - - + - - - + - - - - - < f - - - - + - -·- + - - - - - 1 0 - ~ 4: 5 I N -z >- - - i 0 ! - ' I I • I ! - --- - ·· t-·-- ---1----+-i- I i i ~ - - - --~ - ··+--- + - ~ i I f - - - ·· -- 0 20 60 40 GRID 0V ER VOLTAGE IONIZATI -ON TIMES 8~6 SPEClAL 6S°C \ \S \ --- . ~ ~ t - - - - - J -- t \ l - - - -- 10------;-,- • ~ - + -- - - r--_ II + - - - - - - - + - - - + - - - - + - - - r - --- , ----,.----·- ·- ~-----~- ------11- - + - - - ~ - - - 1 ·..J----t-i - - - - - t - -7 - - - ---- , ·r=-~ ~ +-1 -- - - - - t - - - - - ----_ _____ -- +---- ~--- ~ .............. ___\_.,.__ l _ - - · ·· · · T -s-·R<fear R< , N ~ so-v-""'t-b-_LT_s_._. ~ ~---1~-~-~~·-·- i . , • 1 - -·-t - s ~·- - - - - --- -· ··· - - ~ ~- - -- - •• - 1 '\,~. z ----~ --· 0 - ' --~ +-- i • •• • ----1--r--~- ----- - - - + - - - - i - - - - - - ~ ~-- - - --~- - - + : - - - - - t - - - - ; - - - t - - - - - t -- - - -- ~r---.l • •• • • •• - - •- - -- - - -or-4-----jG.[-_:::_3=:0j:O~V~~~LT~S i 0 20 40 GRID OVERVOLTAGE IONIZATION TIMtS COLD CATHODE TUBES 60 APPElffiIX B "Surge Voltage Breakdown Charaoteristice for Eleotrioal Gaps in 011« By R. w. Sorensen DISCHARGE TUBf FORE-CO~C£.l'HRATION COIL BLOCKING PLATES SWEEP CIRCurr FOCU..SIWG COIL SELECTOR SWITCH TO McLEOD GUACC SWHP PLATfS CAME~A ANO FA~ ADAY CAGE ., ~::;======~==:;~ TO MERCURY VAPOR PVMPS . ' . , ... t ' 000000 CONTROL P.I\Nfl FIGURE OSCILLOGRAPH BENCH BLOCKING PLATES ~ I 400 400 I 400 TRIP . ____IJ 400 GAP I 1000 • O • $00/:\ .t· :2500 900 . (> 900 • • • /' ,¥ • ~50 5"0 • 50 0 .1 MFO (ACH dH - HH H-- H I I SWEEP PLATES 500 ooo ______-_-_-_-11 "'IEG 0 - - L--'-------)--11-+--+---<r--1 10000 10000 BLOCKING VOLTAGE FIGURE 3 0 I ME& BIAS VOLTAGE -SWEEP CIRCUIT ti: - -. -r-4 - ~ n 'i ~ 1 ~ I f 1 fl 1, ' ~-- -H c,_ . - >-- -- ,j ·, .'> .j '7 .... 700 . ---- - - --.~ j k { \ ) • 25 1000 C ~"~ ~.r 12 '2 s '25 :-:::Y.< : : : ~ _ .. J ~ -- - -- ]- l/ 1000 • ✓ __,r 1-., . :> '> ..,..:)> :>: ► > ;> > ,. .? ~ ... R R, , ' '? 'a R, 2000 R, R. . ....~ ~: ' .._ ~ - ·- ,-- • ► - ---· • ,, ' I GOO R, sooo • ~ ' ·- 550 5S0 R. 2000 A., R2 R, ~. R. R,R, - 1000 • ;. :> ,> ,• ;;, L __ - Q ~ • • !::> . --- Ir ,cc;:,r6ac!J . 9'?99c;,9 L__J ..... l. - - - - • --- _J ~ --- --- - --- ---~- FIGURF.. 4 CIRCUITS ON CONTROL PANEL