Color TV, Conversion and Power Tubes

Electron Tubes: A Constant Challenge

When the history of our time is analyzed in future centuries, I feel certain that communications will be singled out as the key factor in making possible the "one world" concept-the basis for man's best hopes for the future. RCA's business is communications, and the people of our Lancaster plant have been a strong link in the development of modern systems through which men ex- change ideas and convey information. Lancaster contributed to the success of the rapidly growing television industry by establishing the first automatic high -volume production facilities for black -and -white picture tubes immediately following World WarII. This was a major milestone in industrial progress. Later, during 1954 to 1961, Lancaster built the first automatic high -volume production facilities to produce shadow -mask color picture tubes. The worldwide industry was skeptical about the practicality of the shadow - mask precision tolerances and complexity. Lan- caster effort proved to be so successful that the H. R. Seelen entire industry has adopted the design and fol- Division Vice President and General Manager lowed RCA's pioneering leadership. Color tele- Television Picture Tube Div. vision was one of the major growth industries of Electronic Components the 1960's. To the engineer goes the credit for the indispensible design of the tube and the tube processing equipment. And yet all that has gone before is only a fore- runner of communications systems still waiting to be developed. Some systems will incorporate computers; some will, perhaps, have solid-state display panels; but without doubt, whatever their nature, the engineers at Lancaster can be expected to play a key development role.

The Lancaster operation is an ultra -modern industrial and engineering complex producing a wide variety of electron tubes for a myriad of exciting and important applications in many fields of modern technology; for example, new medical diagnostic techniques; night vision employinglight from the stars as the only source of illumination; power amplification for linear accel- erators used to probe the secrets of the atom; surveillance satellite cameras used to better identify the natural resources of the world,just to name a few. The advent of solid-state devices and technologies which was reputed to sound the death knell of tubes has, in fact, accelerated the growth of the industrial tube business. Combining the older tube technologies with new solid-state knowledge has advanced the fields C. H. Lane Division Vice President and in which electronics participate and we at Lan- General Manager caster look forward to many new advances in Industrial Tube Division the tube art. Electronic Components At Lancaster, the engineer can pursue his profes- Our Cover The three tubes superimposed above the engi- sional career in an amazing number of fields neering scene on our cover are but a small sam- ranging from mechanical design to new optical ple of the myriad tube types designed and pro- techniques. The Lancaster Plant is unique in the duced by the RCA Lancaster facility. Yet, these the three tubes-a color TV tube, an L -band module, range of professional skills demanded by and an image orthicon-set the theme for this products designed and manufactured here. No issue by representing the three principal areas of where is the engineering challenge so varied or engineeringcompetenceat Lancaster:color, power, and conversion tubes. The engineers work- the opportunity for professional recognition so ing at the image isocon test set are E. M. Mussel - great. The engineer is limited only by his own man (seated) and 0. Choi, both of RCA Lancaster. ambition and aspirations. Photo credit: Tom Cook, RCA Laboratories. Contents

Papers Milestones in color -picture -tube development H. R. Seelen 2 Development of cathodoluminescent phosphors A. L. Smith 5

Gases and getters in color picture tubes Dr. J. C. Turnbull, Dr. J. J. Moscony, A. Month, J. R. Hale 10

Colorimetry and contrast performance of color picture tubes G. M. Ehemann, W. G. Rudy 14

Autotest-automatic production -test and process -control system for color picture tubes 18 W. E. BahIs, F. C. Fryburg, A. C. Grover, J. F. Stewart, N. A. Teixeira, E. D. Wyant Vidicon photoconductors J. F. Heagy 22

The image isocon-an improved image orthicon E. M. Musselman, R. L. VanAsselt 24

Ceramic -metal tetrodes for distributed amplifier service C. E. Doner, W. R. Weyant 29

Klystron for the Stanford two-mile linear accelerator A. C. Grimm, F. G. Hammersand32

Environmental engineering laboratory J. M. Forman, J. B. Grosh36

Design of a 915 -MHz power triode for microwave cooking W. P. Bennett, D. R. Carter, I. E. Martin, F. W. Peterson, J. D. Stabley, D. R. Trout38

RCA Lancaster -25 years of engineering excellence 43

The heat pipe, an unusual thermal device R. A. Freggens, R. C. Turner 44

Ruggedization of camera tubes for space applications J. G. Ziedonis48

The Coaxitron J. A. Eshleman, B. B. Adams 52

Noble -gas -ion lasers R. J. Buzzard, J. A. Powell, J. T. Mark, H. E. Medsger 56

Design of cermolox tubes for single-sideband A. Bazarian 60

A sterilizable and ruggedized vidicon Dr. S. A. Ochs, F. D. Marschka64

Broadband, high -gain, L -band, power amplifier module R. L. Bailey, J. R. Jasinski68

H. R. Seelen

Frequently a company, however heterogeneous its output, becomes identified in the mind of the public with a single, best-known product. The RCA Lancaster plant's life -span has encompassed many developments of high interest to scientists, engineers, the military services, and ordinary laymen. But of all the Lancaster products, the public is most aware of television picture tubes, and, in particular, color tubes. The 19 years during which Lancaster engineers have been engaged in developing and producing color tubes has been an eventful period.

'NROM THE BEGINNING,effort has to aid continued development of the r been applied principally to the RCA colorTVsystem as a whole, and shadow -mask type, which appeared to permit additional demonstrations of most suitable for development of the the system at various stages beginning several forms of color tubes investi- in the Spring of 1950. Ultimately, the gated by the RCA Laboratories. We system developed by RCA served as are often asked, today, why RCA has the "backbone" of the National Tele- adhered steadfastly over the years to vision Systems Committee recommen- this tube. The answer has good tech- dations, which were finally adopted nical and commercial foundations: no by the F.C.C. late in 1953. RCA thus other type has shown so desirable a became the major contributor to for- combination of good performance mulation of the present national stand- characteristics and mass -production ards for colorTVtransmission. suitability, brought about by a long series of engineering developments. Early color tubes Admittedly, the three -beam shadow - In addition to furnishing tubes for the Harry R. Seelen mask system is less efficient than other RCA color program, it was, of course, Division Vice President and General Manager systems as regards light output, and it an important part of Lancaster's as- Television Picture Tube Division has convergence complexities. But its Electronic Components signment to develop the color -tube received the BS in Physics from Providence Col- advantages are dominant: high con- design, with a view to eventual mass lege, and in June, 1968, Mr. Seelen received an trast and resolution, capabilities for production. The approach followed Honorary Doctor of Business Administration degree fromit.Mr. Seelen was appointed Division uniform color fields and high color during this early period beginning in Vice President and General Manager, Television saturation(whichisbuiltinto the 1949 was to explore the potentials of Picture Tube Division, of Electronic Components tube), fine dot -screen structure, and the flat -screen, flat -mask system. Be- on August 20, 1965. Previously, he had been Gen- eral Manager of the division since July 20, 1965. moderate circuit requirements, consid- cause accurate positioning of mask Mr. Seelen joined RCA in 1930 as a tube design ering both complexity and stability. apertures with respect to the phosphor - engineer at the Harrison, New Jersey, plant. Sub- screen dots(register) sequently he became Manager of the Tube Devel- Pioneering work in color Tv had been is extremely im- opment Shop.In 1942, he organized and set up done by RCA prior to World War II, portant, great emphasis was placed on the Engineering organization and laboratories at mounting arrangements for the thin, the new Lancaster plant. In 1943, with the transfer but the major advances in the com- of all non -receiving tube engineering to Lancaster, mercial program occurred after 1945. rather flexible, metal mask. In this con- he assumedchargeof EngineeringServices. An important milestone was passed in nection it was felt that tensioning the Seven years later, he was appointed Manager of all mask in a frame would provide great- Laboratory Engineering at Lancaster. In 1954, Mr. 1949 when field tests began and the Seelen became Manager of the Color Kinescope first demonstrations of the RCA com- est accuracy and stability of aperture Operations Departmentandlaterreturnedto patible system were held for the Fed- position. As a correlative advantage, a Harrison as Manager, Kinescope Operations.In flat screen opened up various possi- 1963, he was appointed Manager, Television Pic - eral Communications Commission. For tube Tube Operations Department ofthe RCA bilities for high-speed phosphor appli- Television Picture Tube Division, until his promo- this early work, the receivers used a separate picture tube for producing cation:initially, silk-screening; later, tion to General Manager of the division. He is a letter -press or offset methods. These Fellow of the IEEE and a Registered Professional each of the three primary colors, and Engineer in the State of New Jersey. He is a mem- the three pictures were combined approaches to fabrication of mask as- ber of the Association for the Advancement of Arts sembly and screen required that all and Sciencies, as well as Sigma Pi Sigma, hon- optically. orary engineering fraternity.In 1955, he received mask -aperture arrays be similar, and the RCA Victor Award ofMerit, the company's The next big step, then, was to provide all phosphor -dot arrays be alike, within highest tribute. a single, directly viewed tri-color tube. the narrow tolerances permitted by the Of prime importance at this time was shadow -mask system's geometry. In Final manuscript received October 10, 1968. the fabrication of a group of such tubes other words, there was a need for

2 "interchangeability"; any mask should degree bulbs were cut near the panel In addition, a new triple -beam gun "fit" any screen. Another advantage of seal, and, after screen deposition and structure, which had originally been using the flat mask -screen assembly mask mounting, were resealed with a developed for the flat -screen tubes, was was the feasibility of adjusting relative low -temperature glass frit. introduced in the 21AXP22. This gun positions of these parts for best reg- included magnetic pole pieces for con- ister, as a final step before insertion These first tests were so encouraging vergence control of individual beams. into the bulbs. Many acceptable tubes that, within the next few months, an In this assembly, the three guns were of the flat mask -screen type were made. intensivedevelopmentaleffortwas tilted slightly toward the axis to pro- The 15GP22 was put in production in being placed on curved -mask tubes. A vide static convergence at the screen 1953, and a larger 19 -inch tube of the 21 -inch size was selected because of the center; dynamic fields, in synchronism popularity of this size in black -and - same type was developed. with the scanning frequencies, could white picture tubes; a round shape then be applied to the pole pieces to was chosen for reasons of stability of Curved screen tubes provide the change in convergence bulb and mask assembly, and a metal angle needed as the beams were de- Meanwhile, various curved -screen, envelope was used for flexibility in flected. This development permitted curved -mask proposals had been con- design. From this development accurate convergence to be obtained sidered. It was fully realized that a evolved the 21AXP22, which went into even at the much wider deflection phosphor screen on the face of the production in the fall of 1954. This angle. bulb would be desirable because of tube incorporated many changes in better appearance, a larger picture in addition to the curvature of mask and the same size bulb, and several other screen. It not only provided a larger The glass tube factors; but it was also realized that screen (260 square inches instead of During and following this period, both interchangeability might be difficult to 88 -1/2 -square -inch size of the 15GP22) round and rectangular glass bulbs were attainif curved masks and screens but did so at a slightly decreased over- investigated to determine whether glass were used.Furthermore, aninter- all length by widening the deflection was potentially desirable for either changeable system seemed more angle from 45 to 70 degrees. technical or commercial reasons, and suitablefor mass production; non - to assess the problems connected with interchangeability would require that Photographic application of the three a rectangular shape. Some of the early each mask be mated to a particular phosphors to the curved face plate glass bulbs using metal flanges have screen, through all of the many proc- involved development of processing already been mentioned. A major im- essing steps. In addition, a problem techniques for production use, as well provement relating to glass bulbs oc- would arise if the curved mask did as design and development of a special curred with the introduction of a not remain sufficiently stable through equipment called a "lighthouse". The special frit glass developed by Corning the heat treatments to which a vacuum purpose of this equipment is to provide Glass Works which devitrifies, during tube must besubjectedduring exposure of the photosensitive phos- sealing of the "top -cap" to the funnel, fabrication. phor layers in arrays of "dots" properly at a temperature low enough not to located so that they are bombarded by harm the phosphor or internal parts. In the fall of 1953, another tube manu- the respective electron beams when facturer announced and demonstrated the tube is completed and in operation. This development was followed by the a tube of the curved -mask, curved - An innovation, called a "radial cor- introduction, in 1957, of the first color screen type. Although this tube evi- rection lens", was made in the light- tube in an all -glass bulb, designated denced register problems, it gave some house to provide a correction by type 21CYP22. For screening this tube, indication that a curved mask might optical simulation of the electron a "degrouping correction lens" was be reasonably stable through tube proc- paths. This compensation is needed introduced. This lens has an asymmet- essing. RCA decided toinvestigate because of the manner in which the ric contour superimposed on the radial this type more thoroughly. deflection field acts on an electron correction contour to compensate for Earlier work had been done in the beam, effectively changing its "source" beam degrouping incidentaltody- RCA Laboratories on a photographic position with increasing deflection namic convergence. This latter innova- method for depositing phosphor dots, angle. tion permitted enlargement of the mask apertures at the center, and increased a method well suited for screen appli- The photographic method of phosphor cation to a curved surface. With this transmission by about 40%. It also per- deposition required development of a mitted a graduated tapering of the background, the Lancaster program mask -frame assembly which could be for design and construction of curved - aperture walls which effectivelyin- removed from its normal position with- creased transmission about 20%. screen color tubes was initiated late in in the front end of the tube or "top - 1953. cap" and accurately replaced several As mask material, cold rolled steel was Among the first of these tubes were times. Stability of the curved mask substituted for the expensive copper - both round and rectangular samples. itself was ensured by mounting on a nickel alloy previously used. The aper- The round tubes used 19 -inch -diameter rigid frame; repeated, precise position- ture walls were tapered in such a way bulbs that had a final closure of welded ing of the mask -frame assembly in the that only "knife edges" could be struck metal flanges presealed to the face - top -cap was attained by special leaf - by the beams. This modification re- panel and the funnel. For the all -glass springs attached to the frame which duced electron scattering and conse- rectangular tubes, standard 21 -inch, 90 - engaged studs on the cap wall. quently improved contrast.

3 The glass bulb proved advantageous new approaches to design and mount- and by yttrium oxysulfide 2 years later. in several ways. Its insulating proper- ing of the mask assembly. Frame The yttrium-oxysulfide red -phosphor ties simplified tube mounting problems weight, replaceability, and thermal efficiency has made the "Unity Current for the receiver designer. Its mechan- stability were among the factors which Ratio" mode of operation a reality in ical stability and uniformity of face had to be reconsidered. A simple, inex- the color picture tube. No longer must contour provided very significant im- pensive, lightweight assembly per- brightness be limited by red -gun provement in register. Moreover, an mitting good beam -to -screen register "blooming". Now all guns may be improved referencing system for the control has been developed for the uniformly driven to their resolution lighthouse and for the frit-sealing oper- entire tube family. The control of reg- capability to provide maximum gun - ation also contributed to consistently ister and contrast has established a screen efficiency and thus a brighter good results. The mask -mounting standard which has been or is now picture. Because of this permitted in- method originally developed for the being emulated by our competitors. crease in current to the screen, the 21AXP22 was well adapted to pro- Important advances have also been mask -frame thermal -compensating system called "Perma-Chrome" has be- vide the accuracy needed forthis made in the art and science of light- system. All of these features which house lens design, aided by use of a come a necessity. Brightness has been provided stability and accuracy con- computer. We have been able to de- gained without sacrificing resolution, purity, or white uniformity. tributed significantly to improved uni- posit phosphor dots to a higher degree formity of both color fields and white, of precision than heretofore thought The success of these tube develop- and to greater ease of setup in the possible. The increase in radial mis- ments has literally echoed around the receiver. These improvements were register brought about by the increase world. Our licensees in Japan and the keys to initial commercial success in deflection angle and the greater non - Western Europe have been tooling up inthefabricationof shadow -mask uniformity caused by factors of panel to make similar tubes, and RCA itself tubes. deformation and panel obliquity have has started up new tube plants in In late 1960, another major improve- required greater control. Also, higher Scranton, Pa., Midland, Ontario, and ment was introduced in the tube type beam degrouping factors brought Skelmersdale, England; in addition, 21FBP22. New green and red (sulfide) about by the increase in deflection our RCA plant in Puerto Rico is sup- phosphors, having higher efficiencies angle and the greater nonuniformity of plying gun assemblies. these factors as a result of yoke char- and shorter persistence,supplanted Color picture devices in the future those formerly used. The advantages acteristics also required higher degrees included an important enhancement in of control. Optimum designs of mask We continue optimistic about the po- picture brightness (about 50%) , free- and lens contours have been developed tential of the shadow -mask tube for dom from "smear" inrapid -action for such control. We have indeed come further development, and we expect to scenes, and better balance between the a long way from the initial simple exploit ideas not yet fully explored. currents required from the three guns radial lens on a path of progress which People frequently ask, "Is the shadow - to produce white. Thus, the all -sulfide has been vital to the success of wide- mask tube likely to be supplanted by phosphor screen permitted driving to angle tubes. some new device in the near future?" Any new device would have to show higher white -light output without seri- Another innovation was the introduc- at least as good performance, and it ous red halo. The improvement in cur- tion of smaller neck size to permit use would probably require novelty (such rent balance also simplified receiver of smaller -diameter yokes. This change setup, particularly the adjustment of increased the efficiency of scanning as a large, thin panel to be hung on the video drive for black -and -white picture wall) ,but far more important than needed for the larger deflection angle, any physical qualities would be the reproduction. In addition, a protective and also reduced beam separation. thus potential for cost reduction. A replace- window of 61% transmissionfilter minimizing misconvergence. Develop- glass was laminated to the tube face ment of the small gun to fit the small ment for the shadow -mask tube which would meet such criteria is not pres- with a clear resin to improve contrast neck required further improvements ently within our field of view; we feel, and color saturation. in beam formation and focus to maintain the quality performance of the therefore, that advanced versions of Today's tubes larger gun used in the round tubes. the shadow -mask tube will be serving us for a number of years to come. The last few years have seen the de- Low -wattage heaters were developed to reduce thermal problems in the stem Gen. David Sarnoff's words at the velopment of the first RCA 90 -degree first tri-color tube demonstration on rectangular tube, the 25AP22, and the area. As a result, movement of parts March 29,1950, still ring true: proliferation of its brothers and smaller because of thermal expansion has been cousins. The rectangular bulb shape, reduced and convergence drift has "Measured in comparison with every and particularly the higher deflection been minimized. Einzel-lens guns have major development in radio and tele- also been developed in the small size vision over the past 50 years, this color angle, have introduced additional prob- which permit fixed focus and thus tube will take its place in the annals of lems requiringa higher degree of television as a revolutionary and epoch- engineering sophistication for satisfac- simplify circuitry in portable receivers. making invention. When historians at tory solutions. Use of the rectangular The development of high -efficiency red - the close of the 20th century evaluate the most important scientific develop- bulb has required solution to prob- emitting phosphors has been extremely ments, I will predict that this tube will lems of bulb stability and alignment, important.The red sulfide phosphor was be among the great inventions of the as well as development of completely succeeded by yttrium vanadate in 1964 second half of this century."

4 Development of cathodoluminescent phosphors

Dr. A. L. Smith

Phosphors are solid materials that have the ability to convert one or more forms of energy into visible or near -visible radiation: luminescence is the generic term for the phenomenon of this conversion. There are a number of excellent books and review articles covering the theoretical aspects of luminescence and the various phosphor systems that have been studied. Some of these treat cathodoluminescence in detail,'21° and one is written with practical commercial development problems as its main theme.' None of them, however, deals specifically with the problems associated with the commercial development of cathodoluminescent materials, the subject of this article.

HOSPHORShave been commercially design parameters which narrows the r important for a much longer period field of investigation to a surprisingly of time than other electronically active few chemical systems. These stringent solids." However, despitethe long design requirements also make the history associated with them and the invention of a new commercial phos- very extensive literature on the subject phor extremely difficult, for not only of luminescence, their theoretical basis must the proposed material have ex- is not nearly as well understood as that ceptional luminescent properties, but of the more recently developed semi- it must also possess the combination conductors, such as transistors. Things of chemical and phsyical properties have not changed much since an intro- which will make it suitable as a screen ductory paper presented in 1954 stated materialfor cathode-ray tubes.Al- that the manufacture of phosphors though characteristics of a phosphor is largely an experimental science, if can be varied within moderate limits, not a craft.' Research, development, there is usually an interaction between and technology of phosphors is today the variables so that as one is im- still a combination of art, intuition, proved,anotherisdegraded. The and fundamental knowledge. result is that any phosphor represents a compromise in which the variable RCA's present interest in luminescence considered most important in a given iscenteredprimarilyoncathodo- application is optimized. luminescent phosphors, i.e., those materials that have the ability to transform the energy of an electron beam Design requirements into visible or ultraviolet radiation, The design requirements of commer- because this energy transformation is cial cathodoluminescent phosphors Dr. A. L. Smith, Ldr. the basis of RCA's extensive cathode- can be divided into two broad cate- Phosphor Development ray -tubebusiness. Color -television gories; those dealing with luminescent Chemical and Physical Laboratory characteristics and those dealing with Television Picture Tube Division tubes provide by far the greatest dollar Electronic Components volume in the tube market; for this physical -chemical properties. The first Lancaster, Pa. reason, phosphor research and devel- considers phosphors simply as energy receivedtheBSinchemistryfromFordham converters and sets their operational University in 1941 mid the MS and PhD from the opment done by the Materials Group PolytechnicInstituteofBrooklynin1943 and at Lancaster is slanted heavily toward requirements in a particular device. 1946 respectively. He joinedthe Chemical and support of this item. The second category views phosphors Physical Laboratory at the RCA Lancaster plart as tube components which must rein 1945. The major part of his career has been in A large number of compounds are the capacity of Engineering Leader of the Phosphor main stable during the manufacturing Development Group. He is the author of a number known to be luminescent, but the process. of papers on phosphors and has been granted specificrequirementthaytheybe fourteen patents related to phosphors and color economically useful in a cathode-ray tube screens. Memoerships include the American Luminescent characteristics Chemical Society, Phi Lambda Upsilon, Sigma Xi device immediately imposes a series of A phosphor must convert the energy and theElectrochemical Society. He has held various positions in the ECS, being Chairman of of an electron beam into emitted ra- the Electronics Division in 1953 and a presidential Final manuscript received October 10, 1968. diation efficiently. This requirement candidate in 1955.

5 may seem obvious,butthereare a number of industrial applications in is known as current saturation, or many phosphors that are efficiently which screens are coupled to a par- "droop -on -drive". excited by ultraviolet radiation but not ticularphotographicfilmtypeor Another equally undesirable charac- by an electron beam and therefore are photosensitive surface. In these uses, teristicis known as color shift, a useless in cathode-ray tubes. On the the output of the phosphor is tailored change of hue of emission with inother hand, there are literally thou- to match, as closely as possible, the creased current. The higher the cur- sands of chemical combinations that response characteristics of the system. rent,the more pronouncedisthe luminesce under electron bombard- Because the phosphors developed by color shift. Color shift manifests itself ment but have an efficiency of conver- RCA are used principally for color in the appearance of a new, spurious electron -beam energyinto isin sionof television, the major interest emission band (usually at lower wave- light that is too low to make them rather narrow regions of the spectrum lengths) superimposed on the desired commercially useful. The most effi- comprising the three primaries used emission band. The phenomenon fre- cient and practical cathodoluminescent in colorTV:red, blue, and green. Col- quently occurs simultaneously with phosphors have been estimatedto orimetry of the phosphors is of para- current saturation of the main band. convert, at best, only 20% of beam mount importance because the gamut Both main -band saturation and color power into radiant energy". Some- of hues obtainable in a color picture shift are detrimental to color picture times the tube designer is willing to is defined by the coordinates of the tube quality because they noticeably accept less than optimum conversion three primaries" ". distort the hue in the highlights of the efficiency if some other characteristic, picture. such as a unique decay property, is In addition to efficient energy conver- his major concern. Under any circum- sion and emission,thepersistence (the phosphorescence or decay rate) Physical -chemical characteristics stance, however, conversion efficiency Commercially, phosphors are applied always has a high priority among the of a phosphor must be of the proper requirements. Unfortunately, there magnitude. The requirements in this to a tube faceplate to form a screen category vary widely. In tubes designed by a variety of application techniques. is little theoretical knowledge to guide Settling, slurrying, and dusting are the the optimization ofthisparameter for industrial or military applications, the range may be from 10' second methods commonly employed. The becausethe mechanism by which method used dictates the average par- beam energy is dissipated in a crystal (or less) to several seconds duration, depending on end use. A medium - ticle size and distribution of the phos- and transformed by a luminescent phors to be used. Ideally, the phosphor center into visible emissionisbut short persistence is desirable for entertainment tube types because the chemist should develop preparation superficially understood. The closing processes that are flexible enough to of this knowledge gap is, today, the decay must be fast enough to insure that there is no smearing of the image produce the variety of particle sizes greatest challenge in the field of phos- required for the different application phor development. An advance in in- in rapid action, yet slow enough so that flicker does not become objection- techniques. This condition is fre- formation in this area could revolu- quently difficult to achieve, however, tionize cathodoluminescence. able.Inmultiple -phosphor screens, such as those used in color television because the high temperatures required Another luminescent characteristic tubes, the persistence of thethree to develop optimum luminescent char- required of a phosphor is that it emit phosphors must be reasonably matched acteristicssimultaneously causes in some predetermined portion of the or color trailing (image blur) becomes crystal growth. The phosphor develop- spectrum. Cathode-ray tubes are used noticeable. mental engineer must therefore make in a wide variety of applications, and suitable compromises to obtain the the phosphors used in them must emit It is desirable that the light output of brightest material that can be applied in an area of the spectrum suitable to each phosphor in a multi -color screen in a reasonable manner. a specific application. vary linearly with power, and each with the same slope, so that high and A phosphor must also be chemically Although most tubes manufactured are low brightness areas of a picture will compatible with the application media. used in the black -and -white or color be properly shaded. Because the total Most application methods involve television entertainment area, there are outliut of the screen is dependent on aqueous systems which are pH sensi- beam power (the product of beam tive. The phosphor must, therefore, be voltage and beam current)itis im- stable in water to the extent that it portant that the voltage and current does not decompose to form basic or characteristic of each of the phosphors acidic constituents. Such constituents be stable. As the voltage of the tube is would alter the properties of the slurry usually fixed, the variation in output formulation and its application charac- characteristics of the phosphor with teristics.Of course,thephosphor variable current becomes the major should be inert so that its intrinsic concern. In some systems, such as sul- luminescent properties are preserved. fides, the light output of the phosphor If the phosphor is not inert,it will does not always increase with current suffer severe loss of efficiency during in the linear manner desired, but be- tube processing. gins to drop off at some intermediate The surface properties of a phosphor operating current. This phenomenon must fit the screening techniques so

Phosphor slurry being dispensed. that the phosphor can be well dis- a rather high unit cost can be tolerated persed in the medium, yet bond well provided some outstanding character- to the glass substrate. A variety of istic can be obtained. The red phos- "coatings," which might be thought of phor now used in RCA's color tubes as a "cement", is designed to facilitate is a case in point. It costs substantially bonding in the particular application more per tube than any previous phos- technique used. Colloidal silica, sili- phor butitsoutstandingefficiency cates, and phosphates are the most makes the added cost worthwhile in frequently used chemical coatings, pre- tube and set performance. sumably because they form surfaces which bond well to glass, the most Group organization and common substrate to which phosphors responsibilities are applied. The laws governing sur- The RCA engineering group devoted face properties are at present not com- to the achievement of the phosphor pletely understood. characteristics described above, the Because phosphors areusedina Materials Group, is located in Lan- vacuum, they must not decompose or caster, Pa.Its activities range from evaporate during processing or under Applied Research through Develop- electron bombardment. Organic lumi- ment and into Pilot Plant production. nescent materials cannot be used When necessary, factory assistance is becauseoftheirinstabilityunder given and close liaison exists between Microscopic screen inspection. these conditions. Their decomposition Laboratory and Factory Engineering products could then poison the elec- personnel. tron -emitting cathode or raise the gas efficiency, persistence, and color - pressuresthatwoulddestroythe Laboratory engineering coordinate measurements. usefulness of the tube. Even when a Laboratory Engineering has the ability material has a vapor pressure low to follow an idea from its basic con- Applied research enough to withstand a vacuum, the ception throughfactory production The worker in the phosphor field is energy of the electron beam itself can and, in addition, makes its staff avail- often accused of alchemy because he cause crystal changes that destroy the able to Marketing and Sales as field sometimesusesratherunorthodox luminescence of the materials. Halide engineers. Laboratory Engineering methods to achieve rather unusual re- phosphors, particularly the fluorides, then is involved in the whole gamut sults. His background should be in have notoriously poor tube life be- of the business enterprise. The general either inorganic or physical chemistry cause the chemical reducing power of philosophy of the Laboratory Engi- and, ideally, should include a very the beam causes permanent crystal neering group isto concentrate on broad spectrum of experience; a sur- damage and ultimately destroys screen phosphor technology andtoleave prising number of disciplines are nec- efficiency. testing and analyses to other groups essary in the successful development more knowledgeable in those areas. of a commercial phosphor. (A sense A phosphor must be able to maintain of humor has also been known to its important physical properties dur- Laboratory Engineering does, how- help.) ingtubeprocessing;i.e.,itmust ever, draw heavily on these supporting demonstrate chemical, thermal, and specialists and the well -instrumented The roleof Applied Researchin vacuum stability. During outgassing, laboratories available to them. phosphor developmentistwofold. a tube is subjected to temperatures of First, its efforts are directed toward Analytical group about 400 to 450°C. At these tempera- the achievement of a basic understand- The Analytical Group performs a wide tures,the organic binders usedin ing of those processes which are used screen deposition are decomposed and variety of services. Of particular note in synthesizing the blue and green - yield a combination of gaseous re- are the spectrographic and X-ray dif- emitting sulfide phosphors. Second, it action products, including CO, CO fraction work which isrelied upon carries on a search for new phosphor and H20. Although the phosphor may heavily in purity and compositional systems and new ways of using older Particlesize beunaffectedbytheseindividual studiesof phosphors. systems. The approach to each type gases at room temperature and atmos- analysis is another function of con- of research is quite different. pheric pressure, the combination of siderable importance. Because some gases and elevated temperature could work involves the interaction of many Efforts in the first area might be called purely scientific in that the approach, prove disastrous. variables, the aid of a statistician is enlisted occasionallyinsetting and at least, can be well organized according to basic principles. Because the Phosphor cost interpreting statistically designed experiments. blue and green -emitters are sulfides One of the more important commerical precipitated from aqueous solutions, considerations not yet mentioned is Colorimetry group such fundamental properties as the phosphor cost. It must be possible to The Colorimetry Group is essential to kinetics of the precipitation and the manufacture a phosphor in a repro- the phosphor development operation, role of the precipitating conditions on ducible manner at a reasonable cost; for it is they who make the necessary the final characteristics of the phos-

7 known to be luminescent, but never temperature are developed bestin testedunder modern cathode-ray this plant. The engineering in the Pilot conditions. Plant is then true chemical engineering as distinguished from the pure chem- Development istry of the other groups. Pilot Plant The job of the chemist who devoleps personnel must have a knowledge of phosphorsisreasonablystraight- production -typeequipment andits forward; he develops new phosphors capabilities, a talent for process sim- andtheirmanufacturingprocesses plification, and a healthy respect for to a fine degree. His interaction with costs.Itis up to them to render a other groupsis much more varied phosphor process practical from the than the research man and includes viewpoint of production. Frequently, an occasional assignment as a tech- the Pilot Plant has been the initial nicalfield representative with Sales production unit for a new material Inserting the shadow mask after phosphor application. and Marketing. (RCA enjoys a very and has sold its product to the tube factory. In the early stages of phosphor phor are being studied. The influence substantial share of the market for development, reasonable cost estimates of the flux and flux systems, including zinc and zinc -cadmium sulfide phos- can be made, scrap potentials dis- phase diagram studies, and diffusion phors usedincathode-raytubes covered and corrected, and process as it affects particle growth, are also throughout the industry.) His knowl- reproducibilityassessed.Interaction under investigation. Much of the work edge must be very broad in the field with other groups is, of course, great- in this first area is conducted along of inorganic physical chemistry be- est with the development engineers classical lines where chemical theory cause he tackles problems as diverse and factory personnel, but some liai- is well developed and mathematical as ultra -purification of materials, high - son with Applied Researchisalso models exist. All theory and models temperature syntheses, control of crys- necessary.Vendorrelationships must of course be interpreted with a tal growth, and surface properties of through Purchasing are also important, view toward fulfilling present needs. materials. It is generally he who must not only in equipment areas, but in Note that the researcher must be con- attempt to satisfy all the criteria listed materials as well. Prior to acquisition versant with a wide variety of chem- in the earlier part of this article, or at of the Pilot Plant, the Phosphor Fac- icalknowledge, for he deals with least determine the optimum tory itself had to do its own scale -up solution chemistry as well as with compromise. and process development. This pro- high -temperature and solid-state chem- cedure was undesirable and led to istry. His life is orderly though com- Pilot plant much delay because tests had to be plex. He seeks new knowledge, but The Pilot Plant has proven to be an more or less within the framework of invaluable aid in the phosphor devel- squeeezd into existing production schedules. existing procedure. opment operation in that it performs a variety of functions including scale- The second area of research, that of up of developmental procedures, rough Technical progress new systems, is quite different. Here, cost analyses, manufacture of phos- TableIis a modified, up -dated ver- intuition and art are as important as phors used in small quantities in either sion of previously published data"'. fundamental knowledge. There is no laborfactory,andinvestigations The listings within each color are guiding theory and much of the work leading to new or improved processes. chronological in time of commercial is empirical. The chemist's work in The equipment used in the Pilot Plant color -television usage and show how this area is frustrating in large part morecloselyapproximatesfactory the improvement in phosphors has because thousands of samples may be sizethan does the lab equipment, contributed to the advances in color prepared before one of even moderate therefore, such items as firing time and tube performance. promise is found. Many of the chemical systems have been worked and Table I-Commercially used phosphors reworked, not only within RCA, but inalarge number oflaboratories Types of emitters Phosphor notation Powder colorimetric data' throughouttheworld.Statistically, x y Y(Lumens/W) Blue the chances of finding a completely Calcium Magnesium Silicate :Titanium CaO: MgO: 2S102: Ti 0.169 0.134 8.7 new phosphor are exceedingly low. Zinc Sulfide: Silver: Chloride= ZnS:Ag:CI 0.1460.052 7.5 Zinc Sulfide: Silver: Chloride, ZnS:Ag:C1 0.1500.059 9.1 The researcher must not only consider Green a compound as such, but he must be Zinc Silicate : Manganese 2ZnO SiO2 : Mn 0.2180.712 31.1 Zinc Cadmium Sulfide: Silver: Chloride' (ZnCd)S : Ag : CI 0.2420.529 56.0 concerned about the way that com- Zinc Cadmium Sulfide: Silver: Chloride' (ZnCd)S Ag : CI 0.3030.587 70.3 pound is synthesized, for often success Red Cadium Borate : Manganese 2Cd0 :B203: Mn 0.6300.370 10.7 or failure is due to some unique pre- Zinc Phosphate: Manganese fl3ZnO: PAN: Mn 0.6740.326 7.0 paratory scheme which imparts just Zinc Selenide : Copper ZnSe :Cu 0.6520.347 17.0 Zinc Cadmium Selenide: Copper: Chloride (ZnCd)Se: Cu: CI 0.662 0.338 11.0 the right amount of crystalline irreg- Zinc Cadmium Sulfide: Silver: Chloride' (ZnCd)S : Ag : CI 0.6630.337 12.6 ularity necessary for an efficient phos- Yttrium Vanadate : Europium YV04:Eu 0.6750.325 9.5 Yttrium Oxysulfide : Europium Y302S Eu 0.6600.340 13.8 phor.Considerationmustalsobe given to the "antique" compounds, I. This notation defines color in accordance with that established by the Commission Internationale de I'Eclairage (C.I.E.). 2. The differences between these two phosphors are in silver content, flux composition, and firing temperature. 8 3. These phosphors differ primarily in their cadmium content. Early phosphors ciency loss of approximately fifty per all respects. has processing stability, The listing under the color red illus- cent occurs. Although a coating capa- shows no color shift on drive, and no trates how RCA's research and devel- ble of slowing decomposition was current saturation. opment activity operates in seeking developed, the close controls required A new process compatible with factory to optimize all design specifications. in both manufacture and screen proc- equipment had to be invented for the The first phosphor on the list -cad- essing forced RCA to drop zinc - production of the oxysulfide (here the mium borate: manganese - satisfied cadmium selenide asa commercial staff of the David Sarnoff Laboratories all requirements except one of the product. helped considerably). Firing condi- most important: quality of emission. tions capable of yielding proper color The emission of the borate phosphor Silver -activated zinc -cadmium sulfide and particle size had to be developed, was much too orange. For that reason A silver -activated zinc cadmium sul- rigid control procedures established, it was relpaced by zinc orthophos- fide of high cadmium content eventu- and impurity levels discovered. The phate: manganese whose color was ally replaced the zinc orthophosphate: Pilot Plant went into production of the ideal. The lower lumens/watt value of manganese, primarily because of its material to establish production rou- the phosphate results from its redder higher conversion efficiency, but also tine, and actually made many hun- color and a somewhat lower intrinsic because its persistence matched that of dreds of pounds of product without conversion efficiency, a severe handi- the other two sulfides more exactly. a single reject lot, to determine reli- cap that led to much research to find a All of the brightness gain indicated in ability, costs, and other commercial replacement. The phosphate was com- Table I was not realized because the considerations. mercially acceptable, however, and sulfide suffered mild degradation dur- was used as RCA's standard red in ing tube processing and showed a Future improvements those early years when color was try- slight color shift on drive, a drawback From RCA's viewpoint, a phosphor is ing to get off the ground. not present in the phosphate. The plus not a commercial success until it has features of the sulfide were considered performed satisfactorily in a market- Zinc selenide: copper to far outweigh its disadvantages, and able cathode-ray tube. To perform sat- The next phosphor to receive consider- for many years it was the standard red isfactorily a phosphor must fulfill two able attention was zinc selenide: cop- component in RCA color tubes. categories of design parameters, one per. This phosphor is almost a classic dealing with intrinsic luminescent example of one whose initial promise Europium -activated yttrium vanadate properties, the other with application was very high but which was subse- The europium -activated yttrium vana- characteristics. Future improvements quently found to lack many of the re- date replaced the sulfide for reasons exclusive to RCA will become increas- quired characteristics. Zinc selenide: not readily apparent from TableI. ingly difficult because intense competi- copper has an outstanding lumens/ Again, it was adopted on the basis of tion has increased not only the number watt value, although its hue is a bit a series of compromises, the sum total of investigators throughout the world, too orange. of which made it superior to the sul- .but the sophistication of the approach fide. One prime disadvantage was its to the discovery of new phosphors. Zinc cadmium selenide: copper cost, some ten times that of the sulfide. An improved red emission is obtained Advantages included no color shift References by addition of cadmium selenide to with high current and no current satu- 1. Garlick, G. F. I.,Luminescent Materials(Ox- ford Univ. Press. New York, 1949). form solid solutions of zinc cadmium ration; persistence was in line with the 2. Leverenz, H. W..Introduction to Lumines- selenide. When this is done, however, green and blue sulfides. The lumens/ cence of Solids(Wiley, New York, 1950). 3. Goldberg,P.,Luminescence ofInorganic the lumens/watt value of the com- watt values shown in Table I appear Solids (AcademicPress, New York. 1966). pound is significantly decreased; the to put the vanadate at a disadvantage 4. Kaltman. H. P.. and Spruch. G. M..Lumi- nescence of Organic and Inorganic Materials percentage decrease is greater than in relation to the sulfide. After process- (Wiley. New York, 1962). ing into tubes, however, the vanadate 5. Kroger, F. A..Some Aspects of the Lumi- should be expected on the basis of nescence of Solids(Elsevier, New York, color change alone. It has been con- shows a very slight improvement over 1948). effi- 6. Curie. D..Luminescence in Crystals(trans- cluded, therefore, that the solid solu- the sulfide, because the sulfide lated by G. F. ). Garlick, Wiley. New York, tion has an intrinsic lower conversion ciency is degraded during processing 1960). 7. Leverenz, H. W.. "Luminescence"Encyclo- efficiency than zinc selenide alone. Al- while the vanadate is not. This is a pedia Britannica(1966). though disappointing in some charac- good example of why the acceptance 8. Palilla, F. C.,Elect. Tech.,Vol. 6 (1968) p. 39. teristics, the zinc cadmium selenide has of a phosphor must be based on its 9. Ouweltjes, T. L..Modern Materials(edited by a final lumens/watt value that is still performance in the tube and not on a B. W. Gonser, Volume 5, Academics Press. N.Y., 1965) pp. 161-257. much higher than the phosphate it was simple powder test. 10. Garlick, G. F.J.,Brit.I. Applied Physics, Vol. 13 (5962) p. 541. designed to replace. A major disad- 11. Larach, S., Shrader, R. E.. Yocum.P. N.. vantage of zinc -cadmium selenide is its Europium -activated yttrium ozysuifide RCA reprints PE -276, PE -280. and PE -291. 12. Henderson, S.T.,"Luminescence," Cam- process instability; in water, at room Shortly after the yttrium vanadate: bridge Symposia,Brit. I. Applied Phys.Vol. temperature, a slow decomposition oc- europium went into production, RCA's 6, Supplement 4 (1955) p. 51. 13. Bril. A. and Klasens, H. A..Philips Res. curs. At the elevated temperatures of research efforts led to the discovery Rpts. Vol. 7(1952) P. 401. tube and screen bakes, the water vapor of europium -activated yttrium oxysul- 14. Bril, A., Klasens, H. A.,Philips Res. Rpts. Vol. 10 (1955) p. 305. evolved during decomposition of the fide, a phosphor with a much higher 15. Bril,A.. Wanmaker, W. L.,Philips Tech. organic binder attacks the phosphor at efficiency than the vanadate. The oxy- Rev.Vol. 27 (1966) p. 22. 16. Hardy, A. E.,IEEE Transactions BTR 11, such an accelerated rate that an effi- sulfide is as stable as the vanadate in No. 2 (1965) p. 33.

9 Gases and getters in color picture tubes

Dr. J. C. Turnbull Dr. J. J. Moscony A. Month and J. R Hale

Picture tubes, like other oxide -cathode electron -tube devices. require a high vacuum for prolonged operatinglife.Unfortunately, the inside tube surfaces become gas sources and destroy the desired vacuum. This problem results from high-energy electron beams striking the tube screen and the shadow mask, in the case of color picture tubes, and liberating various gases from these surfaces. At the same time. high-energy, A. Month back -scattered primary electrons strike and liberate gas from the surface of the con- Color TV Chemical and Physical Laboratory ductive funnel coating. The use of a barium getter in picture tubes forms a film on Electronic Components Lancaster, Pa. the interior tube surfaces, which reacts chemically with stray gases to form stable. received the BS in Chemistry -Physics in 1950 from solid, barium compounds and thus maintain the desired vacuum. This paper dDscribes the City College of New York and continued grad- the use of getters and their function in color picture tubes. uate studies in Physics at the Polytechnic Institute of Brooklyn and CCNY. Prior to joining RCA he was Engineering Leader of the Spectroscopy In- strument Section of the Process and Instruments Company and then the Manager of the Instrument Model and Glassblowing Shop ofEmilGreiner Company. He joined RCA in 1957 as a Spectro- scopist for the Microwave Chemical and Physical Laboratory in Harrison, N.J.In 1959 he assumed responsibilityforAnalytical and Metallurgical Analysis and Physical Testing for the laboratory. He joined theElectronPhysics GroupinLan- caster Color TV Chemical and Physical Labora- toryin1964.AtLancaster he investigated the factors affecting operating of color TV tubes and was responsible for the use of the first high yield exothermic getters in the color TV tubes. Mr. Month received the EC Engineering Award in 1967 for his work in increasing the reliability and life of color picture tubes. Heispresently engagedinthe design and development of future color TV picture tubes. He is a member of the American Chemical Society andAmerican PhysicalSociety; has authored technical papers; and has several patents pending.

J. C. Turnbull J. J. Moscony Color TV Chemical and Physical Laboratory Color TV Chemical and Physical Laboratory Electronic Components Electronic Components Lancaster, Pa. Lancaster, Pa. received the BS inPhysics from M.I.T.in1934 received the BS in Chemistry from St. Joseph's and the PhD in physics from Brown University in College in 1951. He worked with the Electric Stor- 1938. Between 1938 and 1942 he worked on glass age Battery Co. as an analytical chemist, served development and research at the Preston Labora- two years in active duty with the U.S. Army and tories. During World War II. he was employed by then workedfor Waterman Products Co. as a M I.T.,Princeton, and Harvard to work on high Chemical Engineer inthe development of phos- power, high frequency transmitters and modulators. phor screening methods for cathode ray tubes. He He joined Electronic Components in 1945 and from joined RCA, Defense Electronics Products in 1957, then through 1954 was involved in glass and ce- where he assisted in the development of thermo- ramic work in the Chemical and Physical Laboratory. electric materials for refrigeration purposes. In 1958 Thereafter, he was involved with the development he received the MS in Chemistry from St. Joseph's of the metal kinescope and with the development of College.In 1959 he joined RCA Industrial Tube both the metal bulb and the glass bulb'versions of Products, Space Components Engineerirg, where the shadow -mask color kinescopes. Since 1958 he he worked on a variety of topics related to thermionic energy conversion. As arecipientofa has been leader of the Applied Physics Group of J. R. Hale the Chemical and Physical Lab investigating ther- David Sarnoff Fellowship, he attended the Univer- Color TV Chemical and Physical Laboratory mionic cathodes, getters and related phenomena sity of Pennsylvania from 1962 to 1965 where he Electronic Components which affect performance and life of electron tubes. earned a PhD in Synthetic Inorganic Chemistry. He Lancaster, Pa. then joined the Color TV Chemical and Physical received the BS degree from the New York College Laboratory of EC and has been working on mass of Ceramics at Alfred University in 1961. He was spectrometer analysis of residual gases in vacuum employed as a Development Engineer by the Ferro and also on getter and cathode developments. Dr. Company and worked on low temperature glass Moscony has had nine papers published and holds formulations through 1962.Currently, heisen- one patent. Heisa member ofthe American gaged with various glass technology problems, Chemical. American Physical and American Vac- getter design, heater -cathode reliability problems Final Manuscript received. uum Societies. and investigations relative to solder glass technology. Mr. Hale isa member of the American Ce- ramic Society, The National Institute of Ceramic 10 Engineers, The Society of Glass Technology, and The Society of Rheology. THE GETTERS used in black -and - 2 -inch (diameter) neck of the tube. with 185 mg offill. The antenna white picture tubes evolved from Barium fill was 150 mg and actual mounting system needed no changes theearly loop getters which were yield was maintained above 100 mg. to position this getter in the funnel. mounted and flashed onto the neck By observation of the disappearance of This getter flashed better than the 127 - area of the tube above the electron gun. metallic barium. it was estimated that mg endothermic getter that it replaced. These getters were stable in air and 50 mg of barium were typically con- The starting time (time from the appli- used a barium -aluminum alloy which sumed during the first few thousand cation of RF to initiation of the flash) contained approximately 50% of each hours of tube operation. The good per- was 7 ± Is with a total RF application metal. formance of this single getter system time of 30 seconds, and the average The alloy was in the form of a wire has been demonstrated by the excellent yields in production exceeded 150 mg. clad with nickel or stainless steel; the life of the 70° deflection color picture This getter is now being used in the wire was ground down on one side to tubes. smaller color picture tubes. expose the alloy for easy flashing. The Two -getter systems Another vendor produced a large get- getter wire was welded to a metal loop A greatly increased flash area was ter having 240 mg of barium fill. This (see Fig. 1) to facilitate heating by a achieved by using a getter system getter. also shown in Fig.3, has a high -frequency source. This heating mounted in the funnel and directed to number of features which departed causes the barium metal to be flashed flash across the bulb onto the funnel from the conventional ring -type get- from the alloy by evaporation. The and mask areas. Two 130 mg endo- ters. First, it used a greater exposed

ring getter shown inFig. 1 is an thermic getters were used in this area, relative to the weight of fill, than improved version which allows a larger system. As shown in Fig. 2, one was other getters. This was achieved by quantity of barium to be flashed with mounted in the tube neck, and the forming a solid ring of getter material a minimum expenditure of RF energy. other in tube funnel. The system had and pressing the ring into a separate This getter has a stainless -steel channel adequate getter -capacity on the oxygen external retainer ring. This external to which the barium -aluminum alloy capacity test(described later), how- ring also aided the RF heating of the material is pressed. ever, experience showed it to be more getter material. Second, a reflector was difficult to control the flash of two Ring getters can be exothermic or endo- incorporated which became hot during getters and that the flashing produced thermic. The exothermic ring getter flash and re -radiated the barium that higher methane pressure' than with is made by mixing nickel powder with normally deposited on the glass surface a single getter. Although the methane the barium -aluminum alloy powder. as back flash. Third, a ceramic sub- disappeared during tube scanning, it When heating of the getter starts, the strate was substituted in place of metal- caused short-term emission instability nickel and aluminum form an alloy lic supports in the glass funnel around after cathode aging. and release heat to help to bring the the getter support to minimize local- getter quickly to the flash temperature. Large exothermic 'unnel getters lized over -heating and cracking. This The endothermic getter does not RCA's getter -vendors were asked to substrate eliminated glass problems include the nickel -powder addition. develop, to RCA specifications, a and reduced the distance that the single, large getter for the 90° deflec- getter extended above the surface of Getters in shadow -mask tion tubes. An exothermic getter was the glass, resulting in more efficient picture tubes preferred because itis easier to flash coupling to the RF unit. Fourth. the Getters used for shadow -mask color than a large endothermic getter. (Large getter used gas doping with nitrogen. picture tubes evolved along the same endothermic getters can present the This allowed the getter torelease lines as the black -and -white tube ver- problem of liquid aluminum spilling nitrogen just before flashing and, sions. The first color picture tubes from the channel during flash. This thereby, to momentarily increase the (type 21AP22. 70° deflection, large problemisnot present with endo- pressure and confine the flash area. neck tubes) were produced with six thermic getters because a solid nickel - The performance of the 240 -mg getter loop getters mounted above the gun in aluminum matrixis formedduring has been favorable. With gas doping, a circle. These getters had a total max- flash.) A goal of a minimum of 135 it was found that RF generators with imum yield of 90 mg of barium and mg of barium -flash yield was set to frequencies above 450 kHz caused ion- were flashed radially outward with a achieve the same performance as the izationof thegasreleasedduring high -frequency coil onto a limited area two -getter system. Experience with flash and that getter yield was below of the neck. Later, a number of differ- factory flash capabilities indicated that normal. However, when the frequency ent types of getters were used in this the getter fill would have to be in the was reduced to 300 kHz, ionization tube as part of a transition that led to a range of 175 to 185 mg, about twice was not visible. This getter is used in ring getter 36 mm in diameter, located the amount as in the largest exothermic the larger color picture tubes, and has above the gun mount and flashed up- getter of 25 -mm size. provided an average yield of 210 mg. ward and away from it. This getter Based on RCA's specifications, one provided a large area of flash on the Control criteria for getters and vendor developed an exothermic getter getter flash Getters delivered to RCA are checked

ANTENNA MOUNTED for barium fill, barium yield under GETTER controlled flash conditions, and gas

GUN -MOUNTED content. During the inspectionfor GETTER barium fill, the getter material is

Fig. 1-Endothermic getters: Fig. 2-Two-getter sys-Fig. 3-Exothermic getters: 185 - loopgetter(left),ringgetter tem used on the 90° de- mg getter(left), 240 -mg getter (right). flection color picture (right). tube. OXYGEN LINE

OXYGEN INLET Table I.Efficiencies of Various Getter Types CAPILLARY LEAN Getter Tube Barium fill Getter Efficiency Average Oxygen (mg) (1-11/mg Ba) Capacity (l-µ) 70° Tube Endothermic Neck Getter 150 20 2400 90° Tube Endothermic Neck Getter 127 10 4500 Endothermic Funnel Getter* 127 40 Exothermic Funnel Getter 185 40 6200 Exothermic Funnel Getter 240 35 7350 (nitrogen gas doped) Fig. 4-Oxygen capacity test system. Used in combination. broken out of its channel and inspected sion line is used to determine the microns per hour. The quantity of for barium content by chemical anal- amount of remaining barium in a oxygen that causes the cathode current ysis. Barium yield is deter mined by the flashed getter. Each determination is to decrease to 80% of its initial value weight -loss method in which longer the average of two readings, with the is taken as the oxygen absorption ca- than average flashing is used so that getter rotated 90° between readings. pacity (in liter -microns) of the getter. most of the barium is removed from This X-ray technique gives the amount Fig. 5 shows a plot of cathode current the channel. Gas contentisdeter- of barium to within ±3 milligrams as a function of oxygen quantity let mined by an ASTM test,'in which and has beenveryusefulinde- into a new RCA-21FBP22 (70° -deflec- the pressure peak during flash is termining the best flashing techniques tion) color picture tube. The current recorded in a system of known volume. and in the discovery of faulty or poor for all cathodes (red, blue, green)is In production, barium yield must be flashing stations. constant up to 2000 liter -microns controlled and conveniently evaluated. Evaluation of getter performance where it starts to slump. The 80% end The weight -loss method is not wholly point is about 2070 liter -microns for Getter performance is best evaluated all three cathodes. A slower than nor- suitable because the initial weight of in an operating picture tube. The detri- the getter, in an arbitrarily selected mal rate of oxygen admission, 34 liter - mental effect of oxygen on the oxide microns per hour for 60 hours, was production tube, cannot be accurately cathode and the high degree of activity determined. However, because the of barium for oxygen led to a test used to provide sharper end -points on manufacturer maintains control over the curves. The amount of barium method in which oxygen is introduced flashed into the tube was 109 mg; the amount of barium fill in each new into a finished tube. (Oxygen is known getter, the approximate yield can be therefore, the efficiency of the getter in to be present in picture tubes during the neck position of the tube was 19 determined. This method of checking scanning as is carbon dioxide': the ef- liter -microns per mg of barium. yield is used for endothermic getters. fects of these gases can sometimes be For exothermic getters a less time con- seen by oxidation of grid surfaces in Oxygen capacity and getter efficiency suming approach is used, namely, tubes which are operated with insuffi- Ina single getter system, oxygen capac- X-rays are used to perform a residual cient getter yield.) When oxygen is ad- ity is a linear function of barium yield. barium analysis. This is achieved as mitted, cathode current is constant as The getter efficiency (liter -microns of follows: when a primary X-ray beam long as the barium film is active and capacity per mg of barium yield) de- impinges on the face of a getter and its absorbs oxygen. When the film is depends on getter location and design ring, the excited elements in the fill pleted, the oxygen pressure rises and and is sensitive to the distribution of (barium, aluminum, and nickel) emit cathode current slumps. the flash. secondary or fluorescense X-radiation For the purpose of this test method, a For a two -getter system, consisting of of discrete wavelengths. Quantitative a neck getter plus a funnel getter at- analysis by X-ray fluorescence -emis- procedure was developed to break into the tube and admit oxygen gas. In this tached to an antenna spring, oxygen sion spectroscopy is possible because capacity can be expressed by the fol- the intensity of the wavelengths procedure, see Fig. 4, the funnel of a tube is partially drilled with a hollow, lowing linear relation: emitted is proportional to the concen- Q =a N-F-bF tration of that element in the sample. diamond -impregnated tool. Glass tub- In this analysis, the barium Ka emis- ing from a vacuum station is sealed to where Q is getter capacity (liter - the funnel with epoxy resin and the microns), N and F barium quantities

I hole is broken through by impact with (in mg) , and a and b are constants. 0RED

a magnetically controlled rod. Oxygen The chemical equivalent of oxygen re- 1.0 ''-=---1--:--- ". : BGLRUEEEN - is let into the funnel from the manifold acting with barium to form barium through a fixed leak. The latter is a oxide, BaO, is 67.7 liter -microns per o.e section of capillary tube 1 mm in diam- mg of barium. This value is the maxi- eter and 1/4 inch long having a con- mum expected for the constants, a and ductance(at room temperature) of b, for a getter film which reacts com- 0.6 U 0.017 liters of oxygen per second. A pletely with oxygen to form BaO

1 second leak from an oxygen line keeps (assumingthis happens before the 10.4 oxygen pressure in the manifold con- cathode emission isaffected). With stant, usually at 4 microns -a con- this relationship, a least -squares fit of the data taken on many new 25 -inch 0.2 venient value to be checked with McLeod gauge. The rate of gas admis- (90° deflection)color picture tubes sion into the tube is normally 250 liter - gave actual values of a equal to 10 0 1000 2000 0 OXYGEN OUANTITY -LITER MICRON 12 Fig.5-Oxygencapacity test on a 70° tube. EMISSION CURVE SOO

4 and b equal to 40 liter -microns per electron -gun mount structure, making ....".: zoo a mg of barium. Results show that the the partial pressure measurement refer 1 rc to the same gaseous ambient to which i funnel getter with its large area of flash / 0 cipproaches the theoretical value, while the cathode is exposed. The analyser /A 0 sector is attached directly to the neck 100 u the neck getter with its restricted area ,..../4 is less than 1/4 as efficient as the funnel of the tube; therefore, gas measure- ...,PRESSUREOVER - 4i,., GETTER getter. ment can be made during scanning A-

/ Getter efficiencies for the various types operation since one of the three guns 10 in the mount is operable. 0 SOO 1000 MOO 2000 of getters used in color picture tubes OMEN QUANTITY LITER N CRON are listed in Table I. These values are The picture tube mass spectrometer is Fig. 6-Oxygen pressure over getterand cathode current as a function of quantity of a function of flash area. The 70° neck used in a normally processed, sealed - oxygen absorbed. getter has a large area of flash and off color tube which can then be op- to protect the cathode from the test greater efficiency than the 90° neck erated and tested for gas pressures gas and, therefore, its ability to protect getter. The 90° endothermic and exo- throughout a life test. These tests have the cathode during normal tube shown the following: Prior to getter thermic funnel getters have the same operation. flash areas. The 90° exothermic gas - flash, the residual pressure in the tube doped getter, also located in the funnel, is approximately 5 x 10Torr, and When the mass spectrometer is used has a more restricted area of flash and consists mainly of nitrogen, carbon to measure change in oxygen pressure somewhat lower efficiency in terms of monoxide, and hydrogen. Several min- as a function of a constant oxygen -inlet this test than the other 90° getters. utes after getter flash a significant drop rate, and the resulting data is compared to the drop in cathode emission, Getter depletion during tube operation in pressure occurs; the predominant The oxygen capacity test can also be gases are then methane, hydrogen and the results are similar to those shown argon at a total pressure of 5 x 10 in Fig. 6. The color picture tube used used for evaluation of getter depletion for this test was a newly processed in tubes that have been operated. By Torr. Upon subsequent processing, RCA-19EYP22 tube in which 50 mg of use of this procedure, getter depletion e.g., cathode aging, the total pressure barium fill had been evaporated. From is taken as the difference between the continues to fall. Fig. 6, it can be seen that the cathode getter capacity of a used tube and that Results from initial scanning of the emission current decreases to 80% of calculated for a new tube. Through the picture tube show an increase in nitro- the initial value at a pressure of about use of this test method, it has been gen and carbon dioxide pressure; with 2 x 10-` Torr. At this point the oxygen found that oxygen absorption capacity a well flashed getter, however, the com- capacity is 2000 liter -microns. Many of the 2 -getter system in 25 -inch tubes bined pressure of these gases does not similar tests carried out on new was depleted at the average rate of 3 exceed 5 x 10 ° Torr. With continued 19EYP22 tubes and tubes which have liter -microns per hour during the first operation, the pressure in the tube is been operated for thousands of hours 300 hours of normal operation, and at further reduced until after several indicate that the 80% emission level the average rate of 1.5 liter -microns hundred hours the total and partial occurs when the pressure rises to the per hour during the first 1000 hours. pressures do not appear to change ap- range 5 x 10 ' to 5 x 10' Torr. Gas sources within the picture tube preciably. At this point, the total pres- can be evaluated by similar measure- sure, caused chiefly by nitrogen and Concluding .emarks ments on picture tubes which are carbon monoxide, is 5 x 10' Torr. Large barium getters have been devel- modified by omission of tube compo- Thus, the total pressure in the tube oped to satisfy gas absorption require- nents. In one test, the mask, phosphor decreased by a factor of about 10' Torr ments in color picture tubes. Tests screen, and graphite -silicate funnel from immediately following exhaust evolved to assess gettering ability in coating were omitted from a tube, and (before getter flash) until after several production tubes can show whether a an operating tube made by aluminizing hundred hours of operation. getter system is adequate throughout the funnel and face of the bulb, and by The picture -tube mass spectrometer life operation of the tube, and can thus providing small -area patches to allow can also define getter performance. In be used to define the getter system. set up of the picture tube in a receiver. this procedure gas is admitted into the This tube showed no appreciable de- funnel at a constant rate, simulating References 1. Moscony, I.J., Evolution of Methane from pletionofthe neck flashedgetter the gas sources from bulb surfaces. Barium Films (Supplemento at Nuovo Cimento during hundreds of hours of tube oper- The steady-state change in pressure of Vol. V, No. I. p. 7, 1967). 2. Michon, L. and Giorgi. T. A., Studies of Gas ation,thus demonstratingthatthe the gettered gas is determined with the Composition Within Finished Television Pic- omitted components were the impor- mass spectrometer. ture Tubes, (Same p. 104). 3. ASTM Gas Test of Barium Getters (ASTM tant sources of gas. Gettering action is described by the Tentative Proposed Practice for Testing Barium Gas measurement in picture tubes ratio between the gas inlet rate and the Flashed Getters, June 5. 1964, Rev. 6). 4. Collins, R. H. and Turnbull, J. C., Evolution Mass spectrometer analysis of partial pressure change at the cathode. This and Absorption of Gases in Electron Tubes pressures of gas in tubes during their ratio has the same units as pumping (Vacuum 10, p. 27, 1960). 5. Todd, B.J.,Lineweaver,J. L. and Kerr, operating life can define conditions speed (liters per second) , but it is not J. T., Outgassing Caused by Electron Bombard- equal to the pumping speed of the ment of Glass (Journal Applied Physics Vol. that limit cathode performance. A 31. p. 51. 1960). mass spectrometer developed for use getter for the gas under test, unless this 6. Moscony, J.J. and Turnbull, J. C., A Mass Spectrometer for Gas Analysis in Color Televi- in color picture tubes has the ioniza- quantity is very small. The ratio meas sion Picture Tubes. (Supplemento al Nuovo tion source incorporated into the ures directly the ability of the getter Cimento Vol. V. No. 1, p. 93, 1967).

13 Colorimetry and contrast performance of color picture tubes G. M. Ehemann W. G. Rudy

Although manufacturers use many technical terms to describe the quality of color television pictures, the ultimate appraisal is made on a purely subjective basis by the viewer. Therefore,itis more important to produce a picture thatis "pleasing" to most viewers than to strive for an exact reproduction of an original scene which the home audience does not see. A viewer's subjective appraisal of a color picture depends upon the sensations he experiences as a result of various light stimuli. These sensations are produced by many properties of the light which can be measured and George M. Ehemann described in absolute values (e.g., total intensity, spectral energy distribution, hue, Chemical and Physical Laboratory purity, brightness, contrast. and resolution). This paper discusses audience responses Television Picture Tube Division to chromatic stimuli of various hues, and relates the physical properties of these Electronic Components Lancaster, Pa. stimuli to the concept of picture "viewability." received the BE in Engineering Physics from Cor- nell University in 1964. That year he joined RCA's Chemical andPhysicalLaboratoryatLancaster VISIBLE LIGHToccupies the wave - of complementary colors generated by and has been working on color measurement tech- length region between 4000 and the primary triad. Because the proper niques and optical property studies of color television screens. In 1965, while on leave of absence. 7000 angstroms. The spectral energy mixture of complementary colors pro- heearnedthe ME in EngineeringPhysicsat distribution of a light source is a duces white, the straight line connect- Cornell University. detailed description of its color coming each pair passes through the white, position, i. e., the relative intensity of or neutral, point. The location of this its light components distributed along point on the connecting line is an indi- this wavelength region. Each of the cation of the effectiveness of each three phosphors used in color tele- primary color in neutralizing its com- vision is a primary emitter of light in a plementary color. For example, blue separate region of the visible spectrum, light neutralizes its complement, yel- i.e., the three phosphors produce red, low light, very effectively;red and blue, and green light. green are less effective in neutralizing theirrespective complements, cyan Color mixing (blue-green) and purple. (This prop- The triad of red, blue, and green phos- erty is described as relative luminance phors is important in color television of each color, or as the luminance ratio because the proper mixture of these between the complementary pairs.) primary emitters produces white light. The neutralizing power of a primary Addition of any primary color to the color over its complement also affects white mixture produces the hue of a viewer's response to a field of that thatcolor;subtraction induces the color placed in a luminous neutral William G. Rudy complement of the primary color. The (white) background. A neutral back- Chemical and Physical Laboratory primary triad, therefore, can generate Television Picture Tube Division ground of higher luminance (bright- Electronic Components a countless number of colors (the ness) than the color area induces an Lancaster, Pa. totality of which is calledits color received the BS at Dickinson College in 1943 and apparent grayness into the chromatic the MS in Physics at Carnegie-Mellon in 1948 He gamut), as well as white. (Although field. The color at which the threshold has had additional academic work at the American this discussion treats the primary triad of grayness is reached and "fluorence" University atBiarritz,France, Newark College of as pure, spectral, or "rainbow" colors, Engineering, and Franklin & Marshall College.In (the appearance of fluorescence) 1948 he joined the Chemical and Physical Labora- it also applies to the phosphors used in begins to decrease is called a zero -gray tory at RCA Lancaster and achieved senior engi- His major concern has commercial color television, which color.' neering statusin1955. have less sharply peaked spectral beenexperimentalwork embracingconversion, Zero -gray colors for a given hue are cathode ray, and power tubes.His contributions distributions.) resulted in improved performance in camera tubes described by their luminance and leading to four patents, enhanced secondary elec- The standard color mixing diagram purity. The purity of a zero -gray color tron emissioninconversion tubes, and accurate shown in Fig. 1, which was adopted by is the ratio of the luminance of the and meaningful test methods for evaluating thermi- the International Commission on Illu- onic emission, residual gas. and life performance pure or chromatic hue -determining in cathode ray and power tubes. Presently, he is mination in 1931, illustrates the pairs component to the total luminance of workinginthe fieldof colorimetry of color TV picture tube phosphors and has recently published the color mixture(chromatic plus a paper on the chromaticity of narrow -band emit- Final manuscript received October 16, 1968. white additive) which complements ting phosphors.

14 320 00

5,6 GR. EN 70 it in a two -component mix and desat- The uniqueness of the zero -gray colors uratesit.Both can be varied inde- for blue hues lies in the behavior of 60 pendently of the luminous background. 500 required blue chromatic luminance as So a function of color purity in the pres- An experimental technique was used ence of a bright background. The blue 40 to determine zero -gray color curves for chromatic luminance curve in Fig. 3 SLOE -60EEP blue, yellow, green, red, and cyan hues is flat at all color purities above a low SO 490 ;. produced by a P22 television phosphor threshold value; therefore, reduction screen when the grid voltages of the of the white component does not 20 three electron guns were adjusted to 440 require an increase in chromatic illu- 0 generate the proper primary excita- mination, and the zero -gray threshold 3L tions for the standard television white. 46 is maintained all the way to full purity 0 10 20 50 40 SO 60 70 SO For a desired hue of red, blue, or green, as the white component is gradually Fig. 1-Stancard ICI diagram showing spectral colors the grid voltage exciting the corres- reduced to zero. and th3ir corr dements (the color of neutral equals that ponding phosphor primary is increased of a 7 )00` K lack -body radiator). either slightly or noticeably to produce Although a decrease in the white com- P22 PHOSPHOR SCREEN EXCITED low and medium purities, respectively. ponent of luminance in a yellow hue TO DESIRED HUE AND Full purity occurs when the hue -pro- produces a purer yellow on the color ducing primary alone is excited. mixing diagram of Fig.1, for fixed BAFFLE REFLECTING SURROUND background luminance near threshold, ILLUMINATION TOWARDS VIEWER For example, yellow and cyan hues the induction of grayness offsets this are produced when their respective physical gain in purity. As a result, phosphor primary complements (blue additional chromatic yellowillumi- and red) receive less excitation than nation would be required to produce that required to produce white. When the same total luminance of the orig- the blue or red phosphor is unexcited, inal color to a viewer. At zero -gray full purity results in the yellow and threshold, therefore, total luminance cyan hues, respectively. of yellow hues remains constant for purities ranging from zero (completely VIEWER'S ETC Bright light having the color of a white) to unity (pure yellow) . Wheth- IMAGE OF EXCITED 3500°K black -body radiator was used er the "pure yellow" isspectral or PHOSPHOR SCREEN to illuminate a reflecting baffle placed within the television gamut of a broad- Fig. 2-Expe- mental determination o' zero - gray colors. between the observer and the televi- band green is unimportant in matters sion screen; the light passes through concerning fluorence of yellow hues. and shadow areas, proves particularly the viewing apertureinthebaffle So far as fluorence is concerned, phos- useful. without striking the screen, as shown phor efficiency rather than purity of inFig.2. The observer views the emitted color is essential in the green One important factor is the picture screen through the baffle illuminated and red primaries. brightness required for comfortable by a luminous background, adjusts viewing at various ambient illumina- screen luminance until the zero -gray- Picture viewability tion levels. For example, the luminance ness threshold is reached, and records required when pictures are viewed two measurements:thetotallumi- Viewability is best studied independ- with 2.5 foot -lamberts reflected am- nance of the phosphor screen and the ently of a television picture so that the bient illumination, comparable to that luminance of the hue -producing pri- viewer can control some of the physi- encountered in a moderately lighted mary (in the case of yellow hues, the cal properties. When a television pic- room, is approximately 17 foot -lam- ture is simulated with projected color blue -phosphor luminance is recorded). berts. For prolonged viewing, the pic- The first measurement is the zero -gray transparencies, the viewer can manipu- ture is judged uncomfortably bright at color luminance, the second measure- late independently the color, the lumi- 40 foot -lamberts and too dim at 5 foot - ment allows calculation of color nance (brightness) , the resolution, and lamberts. Fig. 4 shows the relation purity. Fig. 3 shows the results of such the ambient illumination to "creole" between thepicture luminance for measurements for blue and yellow the most desirable picture. typical low and high brightness areas hues. and reflected ambient illumination for For each picture, the effects of the satisfactory viewing. Blue colors, which easily neutralize various factors influencing the viewer's their yellow complements, remain flu- judgment can be evaluated from a Although the contrast and the resolu- orent until a high threshold value of pattern produced by substitution of a tion of a color picture are decreased background luminanceisreached. RETMAresolution -chart transparency by increases in ambient illumination, while yellows lose fluorence and ap- for the color transparency. The image viewability can be conserved by in- pear to contain gray at a much lower provides a gray scale, broad black and creased picture luminance. As an ex- threshold. Television -gamut greens and white areas for high contrast, and reso- ample, a picture was set up with a reds behave as the yellow colors, while lution wedges. The gray scale, which reflected ambient illumination of 0.5 cyans exhibit intermediate behavior. has low contrast in both highlight foot-lambert. The white luminance of

15 SURROUND 20

IB 1000

NIGNLIGH 6- YELOW WNITE AREA 4--, 6 ZERO -GRAYS ZERO -GRAYS OF OF BLUE HUE ,..,_-YELLOW HUE 0 14 500 FOOTLAWSERTS 1 SuRRCAPID 110 PER CENT NORMAL vaKei (LUCKIESH AND DAMS) 12 SCENE 00 g 10 INERAGE e - H1016.16876 Q BLUE COMPONENT I

C 10 2 6 SHADOW ARE e- crcyv 0./ 4 s6'7 GRAY SHADOW AREAS 2 GONTAINOW DETAIL 2- PER CENT _FIONIF VISION EeEREMELY lall BILU5, .-v6vIESALN AFC SAM/CONDI 0 06 0.6 02 04 ?-1"*" I COLOR PURITY 2 4 6 6 2 4 6 Ifto 100

Fig. 3-The required luminance of blue and 0 2 3 5 6 7 LUMMARCE-Y007L kW/WATS yellow hues and their respective chromatic REFLECTED AMBIENT LUMINANCE - FOOTLAMBERTS Fig. 5-Behavior of various areas in a scene components at zero -grayness threshold in a Fig. 4-Picture luminance for typical low and judged equally viewable under different am- 500 fL surrounding. high brightness areas as a function of re- bient conditions. theRETMAslide measured 25 foot - flected ambient illumination. lamberts and the highlight contrast placed on the highlights of the picture. This experimental curve nearly coin- be altered if the dim background area ratio was 50. When the ambient illumi- is yellow, red, or green in hue because nation was increased to I foot-lambert, cides with the 50% normal -vision curve of Luckiesh and Moss. This re- of the radical difference between the the contrast ratio was reduced to 26. zero -gray curves for blues and for The viewer then changed the picture sult indicates that the viewability conserved by the viewer is primarily in greens, yellows, and reds in the low - brightness to "recover" the excellent contrast -ratio region, where the fluo- picture. However, instead of increas- the picture shadow areas. rence of the latter hues is less sensitive ing the highlight brightness to 50 foot - There is also agreement between the to contrast ratio. Vertical lines in Fig. lamberts to produce the original con- 5 would then represent constant view - trast ratio of 50, the viewer settled Luckiesh and Moss normal -vision data and the experimental data on induced ability as determined by the zero -gray for a highlight luminance of about 40 condition of the green, yellow, or red foot -lamberts, and a contrast ratio of grayness in a chromatic field. The hue. It is also possible, however, that only 40, as shown in Fig. 5. Luckiesh and Moss curves for 50, 80, 100, and 120% normal vision are not detail in the shadows must be con- The conservation of viewability was associated with independently adjusted served for any hue, or even for true studied by Luckiesh and Moss' using background luminance. In their work, grays; this condition would lead to the black -and -white Snellen eye chart the background is identical to the horizontal lines of constant contrast at various room light levels; their re- highlights. However, when zero -gray ratio at the highlight regions as well sults are shown in Fig. 6. The percent curves for blue and yellow hues in a as in the shadow regions for constant contrast has been recalculated in terms background luminance of 500 foot - viewability of a picture. 3onto of contrast ratio, which is the lumi- lamberts are replotted from Fig. Detail resolution and picture fluorence the Luckiesh and Moss plot, they show nance ratio of the highlights to adja- are not necessarily incompatible, there- cent areas on the eye chart. These significant correspondence. (Color fore, since both are enhanced by gain curves should remain valid for com- purity in Fig.3is converted to contrast in picture luminance. For a scene with parison to color television, in which a ratio in Fig. 5. The contrast ratio for dim, unstructured chromatic back- self -luminous phosphor screen and a color mixtures of chromatic light and ground mixed with ambient illumina- protective gray glass are placed in an white light composing the hues of tion, however, a compromise may be ambient setting. Fig. 3 is the ratio of the total lumi- necessary between increased back- nance of the hue to that of the white The experimental highlight data for ground fluorence and a slight loss in component.) constant viewability shown in Fig. 5 detail resolution in other shadow areas correspond to the 110% normal -vision For dim colored areas such as blue on of the scene as the ambient luminance curve of Luckiesh and Moss in Fig. 6. a projected picture, the zero -gray curve is increased. If picture fluorence were However, the slope of the latter curve for blue shifts to the left in Fig. 5 (and much more critical than detail resolu- indicates that a decrease in highlight produces smaller threshold luminance tion, constant viewability of pictures contrast ratio should not affect the for the same contrast ratio) as a result with green, yellow, and red shadows viewability, while the slope of the ex- of the much lower background lumi- would requireconstantluminance, perimental data indicates picture nance values (0.5 and 1 foot-lambert) and the viewability curves would be degradation with decreased contrast used in the experiment. The experi- vertical or nearly so in the highlight - ratios. This discrepancy was resolved mental curve of constant viewability in area behavior of Fig. 5. Pictures con- when the experimental shadow -area the blue background then agrees fairly taining detail in green, yellow, and red behavior (i.e., contrast ratio and lumi- well with the shifted zero -gray curve shadow areas would not suffer in this nance) was plotted in Fig. 5 for the for blue color areas. These experimen- respect. Because the increase in picture same ambient conditions as those tal results of constant viewability may luminance recovers both fluorence and

16 Lli .2 L2 2.4 drive is quadrupled to produce the 1000 a - 2.2 6 - original average scene luminance of 7 LI(NJ 100 foot -lamberts, the contrast increases 2 4 6 8 10 15 2.0 0 RErt.ECTEDAMBIENTIfL ) - by 35% to 1.8. 0.5 i W LB 2D Table I shows the effect of the face- U plate transmission on contrast for the 16 *.111 -...0111142 e40 a 6- above example, without video -drive BDa 14 "...: 11M0 compensation. Decreasing faceplate 3 )3? 1-1 2- transmission decreases luminance, but 12

increases contrast. 41444111111111111111571PW--

ci i _. In color television, the white light that 2 4 6 8 10 12 14 a 8 mixes with and desaturates a color is AVERAGE SCENE LUMINANCE- FOOTL AMBERTS related to the background luminance Fig. 7 -Effect of reflected ambient illumination on shadcw-area detail contrast fo: two by a factor that involves the television picture elements with luminances L, and 01 6 8 4 6 8 ,0 6 e1a0 faceplate transmission. Zero -gray LUMINANCE -200TLAMEIERTS colors of the various hues are produced 10 Fig. 6 -Constant viewability curves of Luck- iesh and Moss. in the picture shadow areas, and the purity of these threshold colors is con- shadow -area detail(except for daz- trolled by the faceplate transmission. zling background luminances), and In the case of zero -gray colors of yel- because both are desirable in a given low, red, and green hues, varying area of a scene, the viewability is lim- faceplate transmissions also cause ited by shadow detail provided no zero -gray areas of the scene to be re- dimmer unstructured chromatic back- distributed, even when the product of ground exists. the picture signal luminance (video 10 Fig. 7 shows the effects of reflected drive) and glass transmission remains ambient illumination on shadow de- fixed. If glass transmission is decreased tail contrast for a specific case of two to enhance the contrast ratio, the blue adjacent picture elements in a shadow threshold colors return to the original area with luminances L, and L where distribution in the picture after video - drive adjustment. In the case of the 10 100 L, is twice the value of L...(At zero LUMINANCE- FOOTLAMBERTS ambient, the contrast would equal 2.) other hues, video drive must be further Fig. 8 -Effect of glass transmission choice increased until the total picture lumi- for fixed operating conditions as compared The average scene luminance is de- to experimental segment of constant picture fined as the sum of the average luminance is the same as that for the origi- viewability drawn for highlight region. nance of the two picture shadow areas nal high glass transmission. and the reflected ambient. decrease as a result of beam blooming For constant ambient luminance and and beam power requirements will video drive, highlight -area behavior is As an example of the use of Fig. 7, the increase. following conditions are assumed: directlyrelatedto shadow -areabe- havior as the glass transmission varies. Average scene luminance 7 foot -lam- For different faceplate transmissions. References berts (i.e., L, = 4 foot -lamberts); 1. Evans, R. M.. and Swenholt, B. K.. "Chro- Reflected ambient luminance 4 foot - contrast and luminance in the high- matic Strength of Colors:Dominant Wave- lamberts; lights have been compared with the length and Purity," /.of Optical Soc. of Amer. Screen reflectivity 1.0; (Nov 1967). experimentally determined conditions 2. Luckiesh, N., and Moss, F. K.,The Science Glass reflectivity 0.04; for constant viewability. If a particular of Seeing(D. Van Nostrand Co., New York, Faceplate transmission 0.5. scene has dim achromatic detail upon N.Y.. 1943) p. 133. a still dimmer and unstructured blue When no faceplate is used, the con- Table 1 -Contrast and average scene lumi- trast is 1.33. With the faceplate, the background, Fig. 8 shows that 40% nance change as a function of faceplate transmission for a scene of reflected ambient luminance is reduced transmission is a reasonable value for low brightness and low contrast by 75% (doubly attenuated) and the practical use. Values lower than 40%, (Lt/L2 = 2) area luminances L, and L. by 50% including the 24% value for optimum Re- (singly attenuated). The resultis a contrast, reduce picture viewability flected Ave. Face- ambient scene 64% decrease in average scene lumi- primarily because of decreased lumi- plate Lumi- luminance but a 12% increase in contrast. nance. Transmission higher than 40% Trans- nance nance mission (IL) (IL) (IL) Contrast (Although this latter figure seems in- produces a substantial increase in the 1.0 4.0 4.0 7.0 1.33 significant, the eye is capable of dis- reflected ambient illumination and a 0.8 2.56 3.06 4.85 1.37 cerning a 2% shift at this brightness consequent loss in detail contrast. The 0.7 1.96 2.68 3.97 1.40 0.6 1.48 2.30 3.20 1.44 level.) In a television receiver, the loss contrast loss can be recovered by an 0.5 1.08 1.92 2.52 1.47 of brightness can be compensated by increase in the picture brightness with 0.4 0.75 1.53 1.90 t.50 0.3 0.49 1.15 1.35 1.54 an increase of video drive. If the video video drive; however, resolution will 024 0.32 0.92 1.00 1.60

17 W. E. Bahls Autotest automatic production -test F. C. Fryburg A. C. Grover, Jr. and process -control system J. F. Stewart N. A. Teixeira for color picture tubes E. D. Wyant

The automatic production -test and proc- analyses of the data to determine the rier which (through its aging socket ess -control system permits more efficient deficiencies or potential weaknesses in and spring contracts) provides the elec- and more accurate testing of mass-pro- the manufacturing process so that cor- trical interface between the tube and duced color picture tubes than was here- rective action may be taken. Fig.Iis the process equipment. tofore possible. Therefore, this system a functional flow diagram embodying offers the possibility for both improving not only the test capability of the sys- FACTORY AUTOMATIC the quality of the manufactured product tem but also its ability to provide some ELECTRICAL TEST COMPLEX and reducing the cost of the testing control over the processes(exhaust Data Entry Stations procedure. and aging) immediately before testing When the tube leaves the aging area and the analysis and/or reprocessing (Fig. 3) the aging socket is removed AN IMPORTANT STEPin the produc- subsequent to testing. and the test socket is connected. In tion of a complex, mass-produced addition, the data card which travels product such as color picture tubes is PRODUCTION PROCESS with the picture tube is placed in the the final testing. Formerly, color pic- The manufacture of color picture tubes card reader. (As previously men - ture tubes were tested at a number of involves many different steps, includ- test stations, each manned by an oper- ing manufacture of the mask, applica- ator. Because of the many different tion of the screen to the faceplate, tests that had to be made, testing reproduction of the gun assembly, join- Frank C. Fryburg, Program Director quired many people, much test equip- ing of the faceplate to the remainder Color Autotest Project Electronic Components ment, and a considerable amount of of the glass envelope, sealing of the Lancaster, Pa. time, and was relatively expensive. In gun assembly into the glass envelope, received the BS in Physics from Pennsylvania State addition, because of the production evacuation of the tube, and the like University in 1949 and the MBA in Industrial Man- agement from the University of Pennsylvaniain deadlines the number of tests that (Fig. 2) . 1951. He joined RCA in February, 1951 as a Manu- could be performed was limited. facturingTrainee.He has servedinLancaster, Finally, because many people and During the manufacturing process, Pa. as a Parts Manufacturing Engineer, Color Parts Quality Manager, and Color Cost Reduction Admin- pieces of testing equipment were in- each tube is identified by a number. This identification number, and identi- istrator. He is currently an Engineer in TPTD volved, maintenance of calibration was Systems, Quality and Reliability Assurance Engi- fication of certain critical steps in the neering. In the summer of 1964 he was appointed difficult.It sometimes occurred that manufacturing process, are recorded Program Director for the Color Autotest Project. He tubes which should have been rejected is by punching a data-processing card, presently President of the Board of Commis- were not discovered and tubes which sioners in Manheim Twp., Lancaster County and is were perfectly good were rejected. which travels with each tube. For ex- a past National President of the Sigma Pi Fraternity. ample, in the manufacturing process

THE SYSTEM there are a number of different machines employed for evacuating the The system includes five test stations, tubes. The identifying number of the each testing different tube parameters machine and the evacuating cart used and all under the control of a digital for processing a particular tube are re- computer of the stored -program type. corded on the data-processing card for It also includes means ft:Cr automatic that tube. The tube class and the tube indication which shows whether a tube type are also recorded. has passed all tests and, if not, the tests which it has failed and, in some cases, After the above manufacturing steps the reprocessing steps which should be have been completed, tubes of different taken. The computer performs imme- types and classes are transferred from diate(on-line)analyses of the tube different processing equipment and/or parameters which have been measured conveyors to the final processing con- during the testing procedure. In addi- veyor. Each tube undergoes certain tion, it performs subsequent statistical processes which activate the cathode coating of each gun. Such process functions are possible because each tube is Final manuscript received June 27, 1968. located on a separate aging/test car -

18 tioned, this card includes such vital information as the tube class, the tube AUTOMATIC K TO E,Ht UST AGING ELECTRIC MANUA- TEST type, and the identification of the TEST (SUBJECTIVE) equipment used to process the tube). After the card is placed in the card reader and the reader is closed, an RETEST AUTO) LEGEND operator depresses a button on the RE AGE 03J TO) AwPRODUCT DATA reader control panel to indicate the FLOW ANALYSIS PRODUCT test status of the tube. The button de- DATA r DISPOSITION DETECT (AUTO) DATA DEFECT ANALYSIS pressed indicates whether the tube has TAKE OFF 1 \-/ UNALY S IS been tested previously. If all such data \MANUAL SCRAP CONTROL have been properly entered, they will GATE TEST AND MAINTENANCE NET PRODUCTION then be read into the computer. A sig- DECISIONS SCRAP DATA nal is then returned to the card reader L t_ _ AUTO) AUTJ) from the computer indicating that the data have been read and recorded; this signal opens the card -reader mechan- DATA ANALYSIS AND ON-LINE ism. The operator then removes the ",..:NIAIAANRUTAIN,G) PROCESS ADJUST STATISTICS card and returns it to the tube. (MANUAL) (AUTO) In the meantime, the computer is proc- Fig. 1-Functiona system diagram of the automatic production- est and process -control system essing the information read in from the for color picture tubes. data entry station. Some of this in- - FRIT AL- formation is maintained in high-speed I A memory. The balance is stored on magnetic tape for later off-line use. The MOUNT SEAL information retained in high-speed SLEM memory includes the tube class, the tube type, and the position number of 0 the carrier on which the tube is riding. -a E This information is vital to the per- AI formance of the tests in the subse- SPOT KNOCK \- J quent five test stations. HOT SHOT a AGE r N ---) I - J r--- -)

Shorts and Leakage Test Station I I J r I 1 r -dm III ITRANwERI I NLOAD I When a tube reaches this station, an .." SOCKETo r HOOKUP %.,__ ___./..) AREAU 1 I L electrical connection is made between I the tube to be tested and the specific test station. In addition, a signalis r - generated which produces an inter- MANUAL "EST rupt to the computer signifying that i AUTO I L there is a tube in the station ready to TEST CODE test. The computer, having kept track A CARD PUINC-1 0 CARD REA.YER of each of the 120 carriers included in F EMOTE PF INTER the test loop at any given time by means 2-Factory process flow. of an interrupt analysis system, is then TO SUBJECTiV: TEST able to determine the number of the carrier in the station and, therefore, the class and type of the tube to be

tested. With this information, the com- ODOOD r--1 CUTOFF " STATION STATIONPHL puter then determines the tests to be ANNUNCIATION STATION performed on the tube while itis in DATA SHORTS a FOCUS ANODE this station. ENTRY LEAKAGE BREAKDOWN BREAKDOWN STATION STATION STATION STATION It should be noted that once the en- 0 0 0 0 gagement of the contactor blocks of the test carrier and carriage is completed, both the tube and the carriage continue to move along the conveyor. FROM This joint movement continues for a AGING period of 10 seconds which isthe C maximum time available to perform Fig. 3-Factory automatic electr cal test complex. all of the tests on that tube. At the end of the 10 -second period, the contactor circuit of the system. Thus, when arc- 2) The bias voltage is programmed to 0 blocks are disengaged and the test car- ing occurs the current level flowing in by the computer; the measurement device records the emission current, and riage returns to its starting position to the anode lead changes significantly. the computer compares this value pick up the next tube carrier. This change in current is detected and against the predetermined limits. Because of the high-speed switching converted into a digital form which is stored and subsequently transmitted to The emission -current cutoff is deter- capability of this system, the shorts and mined through a series of tests known leakage station has been so designed the computer. Detection of stray emission is accomplished through photo- as search tests. These tests are per- that it is capable of performing a maxi- formed with a fixed negative bias mum of 42 individual tests on each sensitive devices installed in the detector housing which, inturn,is applied to the control grid and a pre- tube. In actual practice, the total test determined value of positive voltage routine has been broken down into 13 placed on the tube face during the entire test period. The photocells sense applied to the accelerating grid. At this group tests and 29 ind:v ;dual tests. point, the cathode current is measured During a group test, voltages are the light which develops as a result of stray emission. and examined in high-speed memory applied to several elements and a which determines whether the next measurement circuit is tied in to one Finally, it should be noted that these voltage step applied to the accelerating additional element. The leakage cur- devices continue to count arcs and de- gird should be more positive (increase rent between that element and those tect stray emission during the entire in current) or less positive (decrease with applied voltages is then meas- test period. The final readings sent to in current) to reach the cutoff point. ured. If this leakage exceeds a prede- the computer constitute a summation This process continues until cutoff is termined value, the computer then of all arcs and stray emission developed reached. This series of steps is per- performs the individual tests. With this during this test cycle. formed three times to check the three method, it is possible to test good tubes guns in the tube. in a minimum amount of time and, Anode -Breakdown Test Station simultaneously, to produce detailed in- This test station is next in sequence A second important test in this station formation on tubes that exhibit leak- and is quite similar to the focus break- measures the gas content of the tube. age. Finally, it is noted that 'this feature down station. The primary difference Inthistest,the search conceptis not only tests the elements in a given is that the over -voltage in this particu- again used;anegative control -grid gun, but can also test similar elements lar case is applied to the anode rather bias is programmed through a series between guns. than the focus element. Therefore, the of steps by the computer until a pre- This station, as well as all other sta- primary source of arcing or stray emis- determined cathode current value is tions, was designed for maximum flex- sion is between the focus and anode achieved. After the search test, two ibility. This system can apply a wide elements of the tube. additional current measurements are made and their values are recorded range of voltages to each element of Experience has shown that although the tube and also read a wide range of in high-speed memory. A subtraction these two test stations appear to be between two values is made and their leakage currents. Such a capability is performing a very similar test,the currently sufficient for measurement of netdifference, usuallyinmicroam- tube -processing characteristics are such peres, constitutes the gas reading. all known color -picture -tube types and that problems can develop in the itis expected that it will handle all anode -breakdown or in the focus - The final test at this station is anode future types being considered for breakdown area independently of each continuity. As the name implies,it production. other. Thus, both of these tests must be determines the continuity from the performed to insure a good quality of anode element of the mount through Focus Breakdown Station the product to be shipped. Finally, it the bulb spacer and graphite coating This stationisthe next one in the should be noted that the system has to the button on the side of the fun- autotest sequence. From the standpoint the capability, through computer con- nel, where the high -voltage circuitry of test -carriage to tube -carrier contact, trol, to provide different over -voltages in the receiver will be eventually con- its operation is similar to the previous for the anode and/or focus element nected. Unless good continuity exists station. In addition, however, it con- as different tube classes are tested. throughoutthisentirecircuit,the tains a sensing device which is placed tube will not work properly and will over the face of the tube. This test Phi Station frequently exhibit a ragged raster. station performs an over -voltage test. This station, although next in the proc- In other words, during the entire test Cutoff Station period that voltages are applied, the ess, is some distance away from the anode -breakdown station. Such a de- is inthe beam current is completely cut off by a This thelastteststation system. There are two major tests at high bias voltage, and excessive voltage lay permits the tube to cool. These is applied to the focus element. The tests are performed at minimum rated this station, each performed at normal purpose of this test is to determine any heater voltage after a short warmup. heater voltages. The first is the emis- The first test in this station is cutoff - instability in the form of arcing or stray sion cutoff test performed on each emission between the accelerating ele- compensated emission current, com- of the three guns. The G, voltage obtained after the successful completion ment and the focus element of the tube. monly referred to as Phi. The test is performed in two parts: of a series of search tests is compared Tube arcing is detected through a spe- 1) The voltages to produce emission - with a predetermined limit.In the cial transducer device in the anode current cutoff are applied to the gun. second test, the bias voltage is pro -

20 grammed to zero and the maximum which stores this information in high- assurance of the long range use of this emission current is read and checked. speed memory. These data identify equipment is provided. For this reason, These tests are then repeated for the with the tube the key manufacturing considerable care in the original sys- other two guns in the tube. Finally, processes used in its production. As tem design and in the subsequent de- the last test in this station measures the tube passes through the autotest sign of the hardware and software was the actual heater current. cycle, pertinent test results are ana- devoted to the idea that long-range lyzed by the computer and recorded flexibility would be maintained. From Annunciation against specific pieces of production the standpoint of hardware design, When a tube has completed ail tests equipment on which the tube was measurement ranges, voltage ranges. in the five test stations, it moves into manufactured. If this analysis reveals and the like, system capabilities were theannunciationstation.For each that a particular piece of manufactur- extended well beyond therequire- tube,thisstationprovidesspecific ing equipment has been producing ments of existing tube types. There- colored tapemarkerstoindicate bad tubes, the computer will display fore,itis anticipated that as future whether a tube has satisfactorily com- this result and the equipment will be tube types are developed, they can be pletedalltestsand,ifnot, what removed from production. Although readily adapted to this system without further processing is required. such condemnation practice has been major modifications in the hardware. in use on a manual basis within pic- At this point, it should be noted that ture -tube plants for many years, itis The software design also included pro- in the data entry station, a 75 -bit anticipated that the additional high- visions for changes. In essence, the section of high-speed memory isset speed capabilities available through system was designed to provide a gen- aside for the tube before it proceeds the autotest system will improve the eralized testroutine. As each new through test stations. This section is effectiveness of the condemnation pro- class of tubes comes into being, a test commonly referred to as the binary cedure and will, thereby, reduce 'the file must be developed for it. This file failure image and during the entire number of faulty tubes generated from can be prepared by personnel familiar autotest cycleit provides the means defective equipment. with testing but with little or no famili- of recording all failures of the tube arity with computer programming. In in question for each of the 75 tests. Data Analysis this manner, it is anticipated that new tube classes can be introduced into the Once a tube reaches the annunciation Another process -control feature of this system without any significantpro- station, the computer scans the binary system is its ability to analyze and sum- gram changes. This fact has been sub- failure image table for that tube to marize certain key variables collected stantiated already through recent de- determine its test status. If the scan- during the test process. In addition, velopment of the Einzel-lens test file ning reveals that the tube has passed the system can provide regular reports all tests, a green label is applied to the of complete test results by the tube foruse withtheAutotestsystem. This development was accomplished tube with the first of the five tape class and by type of defect. Through with no changes in the software. markers.If, however, the tube has these analyses,itis anticipated that failed one or more tests, the computer valuable information will be provided determines the failures in the order to both factory and laboratory engi- SUMMAFY of importance by means of a previously neering personnel which will aid them established priority annunciation table. in their continuing efforts to improve 1 heautomaticproduction -testand tube processing and tube design. process -control system described has With this information, the computer computer -controlled capability for all commands one of the four remaining DESIGN FEATURES functions. This capability allows the labels to be applied to the tube and, control of all five test stations, the per- Accuracy thereby, signifies one of the four al- formance of up to 75 tests on any given ternate processes the tube will under- Analysis of this equipment clearly in- tube, and the testing of many tube RETEST, REAGE, DEFECT ANALYSIS, go: dicates that it is far more accurate than types on successivecarriers. These andSCRAP. any equipment previously used in fac- three features should allow the Marion A remote printer located at the con- tory testing. In many respects, it pro- plant to test all future types and classes veyor unloading pointisconnected vides a level of accuracy superior to of color tubes. In addition, the color to the autotest system. This printer existing laboratory equipment. Two Autotest unit can perform all electrical provides detailed information to the key areas providing a major portion tests of color picture tubes with a tube analyst which indicates the tests ofthis improved accuracy arethe higher degree of accuracy than any that any individual tube failed. Experi- power supplies and the measurement existing equipment. Finally, the proc- ence has shown that these print-outs circuits. In addition, another impor- ess -equipment condemnation and the are proving invaluable for establishing tant feature of this system is its capa- data analysis are expected to yield im- the cause of tubefailure. bility for continuous self -check which provements infactory performance. insures that all test equipment is cali- Thus, with improved process control, PROCESS CONTROL FEATURES brated and functions properly. more acurate test equipment, and a Condemnation more uniform testing facility, the Auto - When the tube reaches the data entry Flexibility test system in the Marion plant pro- station, certain key -process data are A major investment of this magnitude duces a highly reliable and excellent read from the card into the computer cannot be justified unless adequate color picture tube.

21 Vidicon photoconductors J. F. Heagy

Of the many compounds and elements that exhibit photoconductive properties,' three have emerged as most suitable for use in vidicons: antimony trisulfide (Sb2S3), lead oxide (Pb0), and antimony oxysulfide. The first two are used in commercial applications; the third is used in special applications only. This paper briefly describes the development, applications, and properties of these photoconductors.

\A VIDICON CAMERATUBE,the when there is no radiation incident on J. F. Heagy I targetconsists of a layer of semi- the photoconductor.) The antimony- Camera Tube Design conductor material called a photocon- trisulfide photolayers were evaporated Industrial Tube Division Electronic Components ductor. The electricalresistance of in the tube envelope, and exhibited Lancaster, Pa. this photoconductor varies when light, relatively high, nonuniform values of receivedthe BSinPhysics (cum laude) from Franklin and Marshall College in 1957. In 1963, he X-rays, or electrons are incident upon dark current. This was corrected by received the MS in Physics from that institution. it. When the camera -lens shutter is special shading controls in the camera. In1959, after two years of teaching, Mr. Heagy opened for a specified time, a charge was employed by RCA as atype engineerin pattern is formed on the photoconduc- Photoconductor performance was im- Vidicon Tube Manufacturing. Since 1963 he has proved by use of an indium seal that been engaged inphotoconductor development tivelayerthat correspondstothe work in the Camera Tube Design Group. In this image being viewed by the camera. permitted vacuum sealing of the face- assignment,he developedthehighsensitivity The photoconductor is then scanned plate to the glass envelope, or bulb, of photoconductor now used on practically all RCA the tube. This technique allowed the commercial vidicons and customized vidicons. He by an electron beam to transform developed the photoconductor for vidicons used to this pattern into a video signal. photoconductive layer to be prepared monitor radar displays, and was instrumentalin before the faceplate and photocon- the development and improvement of the photoconductor and reticle configuration used in Ranger To perform adequately, a photoconductor were sealed within the vidicon. 8 and 9. Both of these developments resulted in ductor must exhibit adequate sensitiv- Many faceplates would then be coated the granting of U.S. patents. In 1964, he received ity,fast response, adequate spectral in one process, with the result that an RCA Engineering Achievement Award in recognition of his contributions to the advancement of response, low or uniform dark current, photoconductivelayerswere more the vidicon state-of-the-art. Mr. Heagy is a member and chemical stability in a vacuum. uniform than those obtained by evap- of Phi Beta Kappa. In addition, the discharge time of the oration within the tube envelope. In photoconductor, or the time required this pfocess, as many as twenty targets Lead -oxide for it to return to cathode potential, were prepared at one time by passage The lead -oxide photoconductor was must be long compared with the time of the faceplates over an evaporation introduced in 1964.' Fig. 1 shows a between two successivescans. For boat in a long, cylindrical belljar. photograph of the RCA -C23803 lead - commercial Tv applications, the dis- Two passes were made, one in a high oxide vidicon-the Vistacon. Lead - charge time, orRCtime constant, must gas pressure to obtain a porous layer oxide vidicons exhibit extremely low be greater than 1/60 second. and one in a hard vacuum to obtain dark current and fast response, and a hard layer. The resulting vidicons thus are suitable for use in studio Antimony-trisulfide were more sensitivethan previous color cameras. Because these vidicons vidicons because the uniform dark Thefirstcommercialvidicon,the have a very low noise level as com- current exhibited permitted the target RCA -6198, was introduced in March, pared to that of image orthicons under 1952.2 Antimony trisulfide was used voltage to be increased. The RCA - normal studio lighting conditions, and 7038 vidicon which resulted from this as the photoconductor because it had are also simpler to operate, the lead- been discovered that evaporation of process became the standard tube for oxide-vidicon color camera has gained use in film -chain broadcast equipment. Sb2S, at a rather high pressure (10 to wide acceptance in broadcasting. The next improvement in antimony- 1000 microns) in an inert gas, such as lead -oxidelayersundergo argon, produced response to blue trisulfidephotoconductors wasthe Because development of asolid -porous -solid extreme chemical change in a normal light superior to that which could be layer arrangement that increased the room environment, great care must be achieved with any photoconductive sensitivity of the vidicon by a factor taken of the layer after itis evapo- materialpreviouslyused andalso rated. Standard vidicon manufacturing reduced lag and dark current." (Lag of two with no significant increase in lag. As aresult, vidicons could be procedures in which the photocon- referstopersistence,thedelayin used in applications in which insuf- ductor is deposited on a faceplate in changeofphotoconductivitywith ficient lighting conditions had previ- one physical location and the faceplate changes in light intensity measured in ously prohibited their use. However, is assembled to the gun mount in percent of initial value of video signal the response was still too slow for anotherlocationarenotpractical current after illumination is removed because of the possibility of exposure from the photoconductor; dark cur- studio telecasting. rent refers to the current that flows

Final manuscript received October 10. 1968

22 Fig. 1-RCA-CA23083 lead -oxide vidicon, or vistacon. Fig. 2-Decay curve of video signal after illumination is removed from antimony trisulphide IL LUMINATIONUNIFORM OVER PHOTOCONDUCTIVE LAYER 6SCANNED AREA OF PHOTOCONDUCTIVE LAYER 1/2 toy 3/8 INCH photoconductor in RCA -8134 FACEPLATE TEMPERATURES 30°C vidicon (lag is 28%).

many respects electrically. However, A 0 Is PPE1% the basic difference between the two 0,,,,,,p* 52 6 photoconductors isthe barrier -layer O Al Or4.9,4, construction used in the lead oxide #,,.09.12.....,i0.0 P" s,''' z type, which produces a low dark cur- q. rent and a high electric field in the 0 4., photoconductor. The major advan- of the photoconductive layer to air tages of lead oxide targets, therefore, between operations. The photocon- are the extremely low dark current 2 ductive layerisalso susceptible to [ and fast response. Because the dark 0001 2 4 6sap 2 46 110 io gases driven off tube parts and glass current of antimony trisulfide targets 2970 K TUNGSTEN ILLUMINATION ON TUBE FACE -FOOTCANOLES when they are heated during tube is very temperature -dependent,' oper- processing. Consequently, most of the ation of cameras using antimony tri- Fig. 4-Transfer characteristics of a lead oxide and an antimony trisulfide photocon- gun and tube processing must be com- sulfide vidicons may be quite unstable ductor. pleted before the lead -oxide photocon- if the ambient temperature varies as Antimony -ox /sulfide ductor is transferred to the gun mount. much as 5 to 10°C. The dark current of a lead oxide vidicon is relatively Photoconductors other than antimony Because of the chemical delicacy of trisulfide and lead oxide are used in a lead -oxide photoconductor, it is de- unaffected by the same temperature tubes intended for unique applications sirable that the layer be scanned at change; even when the dark current is doubled or tripled, itis still in the in which different scan rates and lag various stages of tube processing to are required. For example, the RCA - evaluateitsphotoconductive prop- nanoampere range and is a very small part of the total signal. C74113 and 4500 vidicons are made erties. Routine vidicon processing with an antimony-oxysulfide photo- steps in which glass or gun parts are The low lag of the lead -oxide photoconductor. This photoconductor can heated may have a severe chemical conductor makes it suitable for tubes provide a sensitivity of 500 µA/lumen, effect on thelead -oxidelayer and used in studio and outdoor broadcasts but also has a lag of 60 to 90% after significantly alter its photoconductive where earlier vidicons exhibited too removal of illumination, and is there- properties. Lead -oxide photoconduc- much lag and smearing of scenes. fore totally unacceptable for broad- tors also tend to undergo significant Lead -oxide vidicons are therefore re- cast use. However, it is quite adequate changes during the first 100 hours of placingthe image orthicons which for slow -scan Tv applications such as scanning;in general, the photocon- were used almost exclusively for live theTIROSandRANGERspacecraft. ductive properties improve. As part color cameras before 1964. of the quality -control procedure, tubes are operated for 100 hours and then Some disadvantages of the lead -oxide References "shelf -aged" for a period of approxi- photoconductors are lack of sensitivity I. Bube, R. H., Photoconductivity of Solids (John to dark reds (6500 to 7000 angstroms) Wiley and Sons, 1960) pp. 281. mately four weeks. If a tube performs 2. Weimer,P.K.,Forgue,S.V., Goodrich, and unity gamma (where gamma is satisfactorily after this period,itis R. R., "The Vidicon Photoconductive Camera the slope of the light transfer curve Tube" Electronics Vol. 23 (May, 1950) pp. classified as a useful tube and will 70-73. shown in Fig. 4). The lack of sensi- 3. Cope, A. D., Goodrich, P. R., Forgue, S. V., haveaprobablescanninglifeof "Properties of Some Photoconductors, Princi- several thousand hours. tivity to dark reds has not proven to pally Antimony Trisulphide" RCA Review Vol. be a major problem inactual use. 12 (1951) pp. 336. However, because of the high gamma 4. Forgue, S. V.,"Porous AntimonyTrisul- Lead -oxide vs. antimony-trisulfide phide," U.S. Patent 2,744,837 filedJune1, value, bright areas of a scene make it 1951 and Zworykin, V. K., Ramberg. E. G., Table I compares the electrical prop- Flory. L. E.. Television in Science and Indus- difficult for the scanning electron beam try (Wiley&Sons, Inc., 1958). erties of antimony-trisulfide and lead - to reduce the signal on the photocon- 5. Dellaan, E. F., "A Photosensitive Device and ItsManufacturing Process," BelgianPatent oxide photoconductors. Measurements ductor to cathode potential. As a re- No. 645,119, March 12, 1964. for antimony trisulfide were made on sult, moving highlights of a scene may 6. Johnson.R.E.,"Vidicon Performancein Extreme Thermal Environments" RCA Review an RCA -8134 vidicon, and those for smear if the electron -beam strength is Vol. 27, No. 3 (Sep 1966) pp. 367-369. lead oxide on an RCA -C23083. Figs. insufficient to discharge the signal. 7. DeHaan, E. F.. Van der Drift, A.. Schampers, 2 and 3 show the decay curve of a D. P. M., "The 'Plumblcon,' A New Televi- The dependence of vidicon sensitivity sion Camera Tube," Philips Tech. Rev. 25, 100-nA-peak videosignalafterre- No. 5 (1963-64). moval of illumination from the 8134 ontargetvoltageismuch greater when antimony trisulfide is used for Table I-Comparison of lead -oxide and anti- and the C23803. Fig. 4 shows the trans- mony-trisulphide photoconductors fer characteristics of both tubes. the photoconductor than when lead oxide is used. The sensitivity of a lead - Photoconductor Antimony Lead Property TrisulphideOxide I Itis evident from Table that the oxide vidicon almost reaches a plateau Operating target voltage (V) 30 50 two photoconductor types differin at 50 volts for a fixed illumination; in Dark current (nA) 35 0.3 Sensitivity (AA/lumen) 200 300 addition, its dark current is relatively Lag (% of initial value of 100 unaffected by target voltage. In an nA signal 1/20 s. after illumination is removed) 28 6 antimony-trisulfide vidicon, the sensi- Amplitude response to a 400 line tivity is proportional to the square of square -wave Tv test pattern at center of picture 25 20 the target voltage, and the dark current Average gamma (slope of is proportional to the cube of the target transfer characteristic) for a signal ouptut current voltage. between 20 and 200 nA 0.65 0.95 Fig. 3-Decay of video signal after illumination is removed from lead -oxide 23 photoconductor in RCA -C23083 vistacon (lag is 6%). The image isocon an improved image orthicon E. M. Musselman R. L. VanAsselt

This paper discusses the basic operation of the image isocon, with emphasis on the electron optics of the reading section which distinguishes the operation of the isocon from the orthicon. The improvement achieved by low -dark -current operation of the isocon is demonstrated by comparison of the image-isocon and image -orthicon performance. In addition, image isocons designed variously for use in studio. X-ray, and low -light level applicationsarebriefly described. R. L. VanAsselt Image Orthicon Product Engineering TI II IMAGE-ORTII ICON CAM IR \ IL IN tered, commercial development of the Industrial Tube Division has been used in broadcast televi- isocon was not continued. In 1958, Electronic Components sion for more than twenty years. It has however, requirements for special ap- Lancaster, Pa. receivedthe BS in PhysicsfromBirmingham - good sensitivity and resolution, and plications revived work on the image Southern College in 1948 and the MS in Physics can televise scenes containing very isocon. During 1960-61 a group at the from Ohio State University in 1949. From 1949 to RCA Laboratories in Princeton, N.J. 1952. he was an Associate Professor of Physics bright objects because of the well- at Birmingham -Southern. From 1952 to1956, he known knee in its transfer character- completed important fundamental was a research physicist with Monsanto Chemical istic. The image orthicon is also work' on isocon operation which pro- Corporation, workingprimarilyininfraredspec- troscopy. Since 1956, Mr. VanAsselt has been with surprisingly rugged, very stablein vided the groundwork for engineering the Industrial Tube Division of RCA at Lancaster, operation and not damaged by inad- development of the tube. The work Pa..principally concernedwithimage orthicon vertent overlighting. The new elec- has continued,' and tubes have been target development. He was instrumentalinthe development of the Mg0 target and the semi -con- tronically conducting glass targets now designed for use in a number of un- ducting glass target. He also played a large part in incorporated in these tubes provide usual television applications-for ex- the development of the 3 -inimageisocon.Mr. long life and freedom from raster burn. ample, astronomical image recording VanAsseltisa member ofPhi Beta Kappa and Sigma Pi Sigma. on a remotely operated telescope'. The image orthicon, however, operates with a high dark -current level that Recent development of the image iso- leads to some undesirable character- con had threebasicobjectivesas istics in the resulting picture quality. follows: Variations in this dark current produce Higher signal-to-noise ratio, partic- background non -uniformity. In addi- ularly in the dark areas of the picture, than that produced byanimage orthi- tion, the dark current adds noise, par- con of equal size and target -mesh ticularly to the signal from low -light spacing; areas of a scene, and this noise limits 2) Retention of the superior resolution, the dynamic range. Finally, beam set- sensitivity, and low -lagcharacteristic ting iscritical because excess beam of the image orthicon; 3) Simplified tube set-up and operating adds directly to the dark current and procedures, similar to those for the im- produces excess noise. age orthicon, so that the image isocon can be readily utilized by persons fa- A recently developed variation of the miliar with image -orthicon cameras. image orthicon, the image isocon, pro- Because the magnetic components are vides performance that far surpasses an integral part of the electron optics E. M. Musselman that of the image orthicon. In par- of the tube, the Camera -Tube Devel- Image Orthicon Product Engineering ticular, the signal-to-noiseratio, opment Group atLancaster co- Industrial Tube Division dynamic range, and background uni- Electronic Components, Lancaster, Pa. operatedcloselywiththeCamera received the BSEE from the University of Pennsyl- formity are improved, while the desir- Advanced Development Group of the vania in 1948. He joined RCA Lancaster in 1948 as able characteristics of the image Commercial Electronic Systems Divi- a manufacturing engineer. He was promoted to orthicon are retained. Manager, Camera TubeFactoryEngineeringin sion at Camden to develop the com- 1956.In1959. he was made Manager, Develop- bination of tube plus focus and de- mental Types and Facilities for the Pickup Tube Image-isocon history Factory. Mr. Musselman was assigned to Storage flection components. Tube Product Development in 1960. In this capacity The image isocon was invented by P.K. he was responsible for the design of scan con- Weimer' in 1947. Because the electron - Image-isocon operation verters.including transmission type.participated in the development of the 7 -in display storage tube optical requirements for isocon opera- Fig.1 shows a cross section of a typi- and designed a compact 5 -in display storage tube. tion were more sophisticated than for He is co -holder of a US Patent for "Insulated Mesh cal RCA image-isocon system. The Assembly for Vidicons", and also holds a patent on image -orthicon operation and prob- outer coil structure and the faceplate an improved target support structure for pickup lems with tube set-up were encoun- coil provide a uniform axial magnetic tubes. Most recently he served as project engineer on the development of a ruggedized 3 -in image field over the entire length of the tube. isocon. Mr. Musselman is a member of Eta Kappa Final manuscript received October 10, 1968. Motion of electrons inthe tubeis Nu, Tau Beta Pi, Sigma Tau and IEEE.

24 lit -STAGE TARGET ELECTRON MULTIPLIER TARGET HORIZONTAL AND VERTICAL STEERING HORIZONTAL AND VERTICAL GRID No 4 RELATES ALIGNMENT COIL most easily discussed in terms of three PHOTOCATHODE, GRID No 6 DEFLECTING COILS

FOCUSING COS. , regions: the image section, the target/ FACEPLATE FOCUSING scanning -beam section, and the separa- COIL N:411:ei::::1XV.::**::::::::::!X11511$ tion section. CAMERA LENS .::::::::):<::*:e.rMe.OrdeOrie :44:29:0

Image section ! I Operation of the image sectionis 7 TELEVISED identical to that of an image orthicon. SCENE When an optical image is formed on E.- veSN SCATTERED , REFLECTED AUXILIARY \- ELECTRON GUN ISP No 5' RETURN BEAM RETURN SEAM ALIGNMENT COIL the photocathode, photoelectrons from ELECTROSTATIC each point of the image are accelerated PRIMARY REAM AL GNMENT PLATES IMAGE SEPARATION - DYNODE SCANNING SECTION toward the target and focussed by the SECTION SECTION axial magnetic field. These photoelectrons arrive with sufficient energy to Fig. 1-Schematic diagram of image isocon and associated magnetic components.

produce secondary electrons and i ++ (1) Rearranging Eq. 1 and substituting charge the target positively. The secon- where I is the primary current from in Eq. 2 yields the following equation dary electrons are collected by a fine- for output current of the orthicon: metal mesh close to the target on the gun to target, i,is the current con- side facing the photocathode. The tar- ducted into the target, i,is the scat- = M (I - (3) getisa thin glass membrane with tered electron current, and i,is the Because the primary beam current 4 reflected electron current. critical physical properties. It must be isconstant, the signal current is a sufficiently thin and resistive so that In both the image orthicon and image function only of the conduction elec- the charge pattern produced by the isocon,the scattered electrons and tron current. photoelectrons is not degraded signif- reflected electrons constitute the return The polarity is negative and the signal icantly during a frame time by lateral beam. In the image orthicon the video leakage. is riding on top of a large dark current, signalisobtained by directing the which adds noise. The isocon signal, The charge pattern on the targetis entire return beam into the multiplier. on the other hand, is derived from the detected by the scanning beam, which The output current can be represented scattered electron current. No dark deposits enough electrons on the tar- as the sum of the reflected and scat- current is present (if perfect separation get to neutralize the charge acquired tered electron currents times the gain is assumed) and the noise is related during each frame. The time constant of the multiplier, as follows: only to the amount of signal current. for the conduction of this charge through the target should not be much = + i.) (2) Fig. 3b shows the advantage of the longer than a frame time; this require- isocon when excess primary beam In the isocon, the return beam is spilt current is used. Although the orthicon ment sets an upper limitto target and only scattered electrons are per- resistivity. The mechanism for extract- signal remains the same, the dark cur- mitted to enter the multiplier. (The rent and noise increase. The isocon ing the signal from the scanning beam direct relationship between scattered - is discussed later. signal and noise, however, remain un- electron current and conduction - changed because the excess reflected electron currentisutilized and the Target/scanning-beam section electron current is separated from the scattered electrons are used to form the scattered electrons before they enter The image isocon differs essentially video signal.) As will be shown later, the multiplier. from the image orthicon in the target/ the separation of scattered electrons scanning -beam section and the separa- and conduction electrons results in a Separation section tion section. To understand this dif- portion of the scattered electron cur- ference, it is necessary to describe what rent being lost. For the same image - Separation of thescattered -electron happens in both an image orthicon and sectionconstruction and multiplier beam from the reflected -electron beam an image isocon when the primary gain, however, the isocon signal cur- is achieved simply by placement of a beam strikes the target. As shown in rent is slightly greater than the signal hole in the first dynode of the electron Fig. 2, when the electrons in the pri- current from an image orthicon. multiplier, as shown schematically in mary beam approach the target, three Fig. 1, and with greater detail in Fig. events occur: Fig. 3a compares the signals from the 4. As shown. the beam of reflected image isocon and the image orthicon electrons is always contained within 1) Some electrons enter the target and when the primary beam is adjusted to the envelope of the scattered -electron neutralize positive charges in the stored picture. just neutralize the charged target. A beam. The electron optics of the read- 2) Some electrons strike the target and few gray -scale chips are shown to- ing section is designed so that, when are scattered. gether with the electron currents gen- the composite beam strikes the first 3) The remaining electrons do not quite erated at the target as the beam scans a dynode, the reflected electron beam reach the target and are specularly line through this pattern. As discussed dis- reflected. passes through the hole and is previously, the orthicon signal is com- carded, while a large fraction of the The current distribution at the target posed of both scattered electrons and scattered -electron beam strikes the satisfies the node equation: reflected electrons. first dynode and generates secondary

25 SCATTERED ELECTRONS w

REFLECTED ELECTRONS

Fig. 2-Current distribution attarget.

-1_F_J-r-SAAL

StAL DARK Fig. 4-Idealized trajectories of image-isocon beams. CURRENT DARK CURRENT the gun. The primary beam then signal. For an isocon to work properly, _L MADE ORTHICON IMAGE ORTHICON spirals toward the target, and the re- this residual scan must be very small, flected -electron beam spirals away approximately 0.01 cm or less. A prac-

fT5-1-1_SIGNAL -1-1-L_SIGNAL from the target. The scattered -electron tical solution was obtained by design MADE ISOGON DARK IMAGE ISOGON DARK beam, however, leaves the target of a deflection yoke that, in combina- CURRENT CURRENT tion with the new RCA image-isocon ( A ) BEAM SET TO JUST (D) BEAM SET WITH 50% aligned with the axial magnetic field. DISCHARGE TARGET EXCESS CURRENT The primary and reflected beams still electrode configuration, provides uni- Fig. 3-Target current as a function of light level and beam setting. cross at the antinode, but the spot is form landing of the beam across the displaced from the center of the scat- target and essentially zero residual electrons which are directed into the tered beam. Fig. 5b shows the cross- return -beam scan. This yokeisan second stage of the electron multiplier. section of the return beam for the same important part of the system of tube Although the concept is quite simple, light levels when transverse energy of plus deflection and focus components the return beam must have certain 0.5 eV has been applied to the primary needed to generate good isocon critical characteristics for proper sep- beam. The low -transverse -energy scat- pictures. aration to occur: tered electrons which represent the 1) The diameter of the scattered -elec- low -illumination signals are spread out Tube Alignment and Adjustment tron beam (at an antinode )must be over a large diametet at the antinode, Proper operation of the image isocon large as compared with that of the re- while the diameter of the reflected flected -electron beam. requires that the electron beam be ac- 2) The return beam must not contain beam is unchanged. Scattered electrons curately directed to the target and back any residual scan. can therefore be captured from all through the separation aperture. The To fulfill the first requirement, itis light levels, and a linear gray scale is primary beam is started on a helical necessary to add some transverse obtained. path and directed through the separa- energy to the primary beam. Fig. 5 It would appear that if greater trans- tion aperture by the combined effect shows the cross section of the return of the alignment plates and alignment beam at an antinode for various light verse energy were given to the primary beam, greater ease of separation would coils (shown in Fig. 1). The align- levels within a scene. In Fig. 5a, the ment -plate differential voltage is preset primary beam is aligned with the mag- result. However, excess transverse energy increases the axial velocity dis- to a value specified for the tube, and netic focus field and the target is per- the alignment -coilcurrents are ad- pendicular to the field. At low illumi- tribution of the beam electrons as well as the radial velocity distribution, and justed to center the primary beam in nation levels, as shown by the extreme the aperture. These adjustments of the left-hand representation, the maximum the beam becomes less effective in dis- charging the target, causing lag. primary beam correspond to the center transverse energy of the scattered elec- of the range for which video signal trons is small and the diameter of the The second requirement for proper is obtained and cause the primary scattered electron beam is essentially separation of scattered electrons from beam to spiral out from the center of equal to that of the reflected -electron reflected electrons is that the reflected the separation aperture toward the tar- beam. As a result, a condition called beam must always return to the same get. The reflected beam spirals back, "black clipping" occurs in which both point in the separation plane regard- as shown in Fig. 4, and criss-crosses the scattered electrons and reflected less of what part of the target is being the primary beam through the same electrons pass through the hole in the scanned. In image -orthicon cameras, a area at an antinode in an ideal system. separation electrode. small amount of residual scan' is pres- Thus, the reflected beam should con- This condition is corrected by adding a ent at the first dynode of the multi- tinue back toward the multiplier and small amount of transverse energy to plier and information about the surface exit through the center of the separa- the primary beam just after it leaves of this dynode is present in the video tion aperture. The image isocon has

26 0AO RATIO 50/ I VOL TGE ABOA 0 0e CUTOFF Iv I 60/1 PERIMETER Of REELECTED ELECTRON 'EAT 35 TERMETER OF SCATTERED .50/1 - -0/1 Fig. 6-Signal-to-noise so-30/1 ratio as a fur ct on of light level III PRIMARY BEAR /RANSTERSE ENERGY 10*y within scene. .20/1 25

10/1 7 20 = ;*, e

15) PRIMARY SEAM TRANSVERSE ENERGY T 55, 15-5/1 Fig. 5-Return-beam cross section at an anti - node as a function of target -oltage and transverse energy of primary bean. 7 ,,5-3 5 3 S 5 7 83-2 2 5 5 IL LUMINAT ION ON A10TOCATOODE FOOTCANOLES steering plates, also shown in Fig. 1, light level. In the isocon signal, how- Signal uniformity produced by the new to direct the reflected beam exactly ever, the noise decreases with signal image isocons is exceptional. The scan through the center of the separation level and the signal-to-noise ratio optic developed to reduce residual scan aperture. The use of these steering varies approximately as the one-half in the return beam also provides uni- plates assures good separation while power of the light level. The curves form landing of the beam with less permittingreasonabletolerancesin show that the dynamic range of the than 0.1 -volt variation across the tar- tube construction and positioning of isocon is at least ten times greater than get. The absence of scanning at the the tube in the yokes. that of the image orthicon. This wide separation aperture guarantees uni- The adjustment of the steering plates dynamic range of the image isocon also form separation. allows the camera to move from scene is made in two steps: to scene without attention to the beam - Background uniformityisexcellent 1) The correct polarity of signal is ob- and dynode problems are eliminated. tained by a coarse adjustment of the current setting used. steering -plate differential voltages.(If The positive -polarity, low -dark -current the reflected beam strikes the first The resolution and the amplitude re- read-out makes variations in gain dynode a negative picture will be across thefirst dynode surface less obtained.) sponse of the new image isocon compares very favorably with the best critical. Also, the first dynode of the 2) The target voltage is reduced close isoconis located at an antinode at to cutoff, and any bright edges are elim- image orthicons operating in well de- inated by fine adjustment of the steer- signed yokes. This excellent perfor- which the current density is a mini- ing -plate voltages. mance results from the electron optical mum and the surface is not in focus. The bright edges result when reflected design in which the magnetic compo- Finally, the absence of scanning at the first dynode means that no video infor- electrons are deflected at the target by nentsandthetubestructureare sharp black -white transitions and ac- properly matched. Image isocons typi- mation can be developed. Dynode quire additional transverse energy.' At cally provide response of 80% at 400 problems have not been observed with an antinode, these electrons are located the new image isocons. TVlines without aperture correction. just outside the perimeter of the re- Corner response shows less than 20% Incomplete target dischargein one flected beam and can become a part dropoff and no dynamic correction is of the signal. The separation aperture sweep of the scanning beam results required. Limiting resolution is 1200 in of the image isocon is large enough to a residual signal characteristic TVlines at an illumination of 0.02 foot called lag. The RCA Dev. No.C21093, allow these deflected electrons to pass candle on the faceplate. Picture geom- an image isocon with a high -capaci- through and be discarded with the etry is excellent. reflected beam provided the reflected tance target -to -collector -mesh assembly, displays residual signals less than 10% beam is accurately centered. The elec- Slightly lower amplitude response re- trons striking the first dynode are then of peak amplitude during the third sults from use of fiber-optic faceplates scanning field when operating at a light only scattered electrons and an isocon and/or lower -capacitance target mesh level with highlights at the knee. The signal free of exaggerated edgesis assemblies. Fig. 7 compares amplitude low -capacitance target of a newer type, generated. response for image isocon and image the RCA Dev. No.C21095, offers fur- orthicon. ther improvement in dynamic re- Image-isocon performance sponse. This performance is achieved Low dark noise is an outstanding fea- High -resolution, slow -scan versions of when the tube is properly set up, i.e. ture of the new image isocon. Fig. 6 the new image isocon have been made with minimum transverse energy re- compares the signal-to-noise perform- by increasing the glass -target resistivity quired for good gray scale given to the ance of an image isocon with that (up to a factor of 10' above the range primary beam. This low -lag perform- of an equivalent image orthicon as a used for 30-frame/second television). ance results in excellent motion - function of the light available in a Lateral leakage is minimized and long capturing capability similar to that scene. Because the noise in the image - storage times are achieved without obtained with the image orthicon. orthicon signal is nearly constant for special cooling equipment. Applica- all light levels(the beam setting is tions requiring long integration time Application not changed during scanning of a given and good resolving power, such as The image isocon is a versatile camera scene) ,the signal-to-noise ratio is ap- astronomical observation, find the new tube. The specific requirements for a proximately a linear function of the isocon a useful tool. particular application can be met dur-

27 the process frequently employs "black living subject with a minimum of radi- 100 ALL rusts 01.7.61THEOZIArxzt,rf(j!k7,7 90 ation dosage. The requirements for the T,...,('7:Er stretch"-a gain nonlinearity designed HK 50 - - - RCA DIV NA 031055 IMAGE ISOGON to correct for an unsuitable gamma tube are similar to those for the low - 41 70 characteristic in the display device at available -light application: a real-time AO RCA DCVNA C21099 IMAGE ISOGON the output end of the system. The low target performance and good response 50_RCA 7190 IMAGE 21 90 ORTHICON dark current and low noise of the to low -contrast dynamic -scene RCA 11,0 30 IMAGE image isocon make it uniquely suited material. ORTHICON 20 to this form of signal processing.) Conclusions 200 400 600 500 1000 1200 1100 Signal output of the new image isocon Tv LINE NUMBER isdirectly proportional to lightre- The new RCA image isocon with its Fig. 7-Amplitude response as a function of associated components produces pic- TV line number. ceived at the photocathode (unity gamma factor) . The gray scale is there- tures that are a distinct improvement fore linear and provides excellent con- over those obtained with image orthi- ILLUMINATION 2050 RCA DEv C21093 cons. Performance features of the INIGIMR-CAPACITY TARGE T1 trast between steps of varying illumi- KNEE nation at all levels below saturation, as isocon include the following: RC DEV No C21015 I COWER- CAPACITY TARGET) KNEE indicated in Fig. 8. 1) Improved signal-to-noise ratio, particularly in the dark areas of a scene, 11, A spectral response emphasizing the which greatly increases dynamic range visible portion of the spectrum can be and allows gamma correction without - obtained with this system by employ- degradation; ingaconventional optical -glass 2) Excellent resolution both in limiting faceplate in conjunction with a photo- value (1200TVlines) and in amplitude 001 1 I 1 1111E I 1 1 1 1111 1 1 1 11111 1 1 1 I I I response; 2 cathode such as the potassium -cesium 10's 2 "I0-4 2 " 10 3 0'2 3) Picture background free of dynode FAWN. AYE ILLUMINATION IFLI-L en/ FI2 antimonide (bi-alkali) or the sodium - defects and shading; 4) Fast and easy setup procedure com- Fig. 8-Signal current as a function of light potassium -cesium antimonide (multi - on photocathode. alkali) types. Response of these photo - parable to that for image orthicon surfaces is shown in Fig 9. The tube operation but without the critical beam - setting problem associated with image 0 can also be made to respond to radia- orthicons; tion in the ultra -violet region of the 5) Linear transfer characteristic and spectrum by use of suitable face -plate low lag attained simultaneously by use

iALKA1.1) materials, such as quartz or lithium of optimum electron -beam transverse PHOTOCATHODE energy; fluoride, and a vv sensitive photo- 6) Very good picture geometry and uni- 001 cathode. form resolution. For extremely -low -available -light situ- The image isocon has inherited many INALPALI ations, such as military surveillance, a PNOTOCATNOOC good features from a long image - multi -alkali photocathode is preferred orthicon tube -development program. in the camera tube because of its broad The burn -resisting and long -life elec- 0 001- - 5000 000 5000 6000 7000 6000 5000 spectral range, which extends into the tron -conducting glasstargetsdevel- WAVELENGTH - infra -red region. The tube uses a re- Fig. 9-Photocathode spectral response oped for image orthicons are equally duced target -to -collector -mesh capaci- effective in the image isocon. The good ing manufacturing by adjustment or tance to minimize smearing and mechanical and electron -optical fea- change of one or more of the following therefore increase the dynamic resolu- tures of the image mount developed factors: target resistivity, target -to - tion of images reproduced from fast- for type 8673 image orthicons have collector -mesh capacitance, photocath- moving dimly lit scenes. If additional also been applied to the image isocon. ode formulation, or face -plate ma- light amplifiers or spectrum converters Mechanical development is continuing terials. are then required, a fibre -optic face- on the isocon and should provide In a typical television -broadcast appli- plate can be used to provide the optical sturdy structures capable of withstand- cation, for example, where thereis coupling. The electron -conduction ing the severe shock and vibration of ample and controlled studio lighting, glass target, proven in image orthicons, rocket launching. the target of the camera tube is read- has been utilized to provide long life, out repetitively every 1/30 second. This resistance to scanning burn, and resis- References rate sets an appropriate value of target tance to highlight damage. Scan zoom I. Weimer. P. K.. "The Image Isocon-An Ex- can be employed without damage to perimental Television Pickup Tube Based on resistivity at 10" ohm -cm. A maximum the Scattering of Low -Velocity Electrons" RCA the target. Three -to -one ratios are read- Review Volume 10 (1949) p. 366-386. target -to -collector -mesh capacitance, 2. Cope, A. D., Peterson, C. C.. and Borkan, H., therefore, produces the desirable pro- ily achieved by appropriate reduction private correspondence (1961) p. 36.37. of the scanning power. 3. Cope, A. D., and Borkan, H., "Isocon Scan- nounced knee characteristic, as well as A Low Noise, Wide Dynamic -Range Camera a very high signal current under the An interesting application of the image Tube Scanning Technique" Appi. Optics. Vol- ume 2 (March 1963) p. 253-261. ample lighting conditions. This high isocon is its use in medical X-ray equip- 4. Cope, A.D., and Luedicke,E., "Camera signal current, in turn, requires mini- ment. The intrinsically high sensitivity Tubes for Recording Stratoscope 11 Telescope Images" RCA Review Volume 27 (1966)p. mum amplification in its subsequent and resolution of the isocon permits 41-56. processing.(It should be noted that continuous detailed observation of the 5. Weimer, P. K., loc. cll. p. 376.

28 Ceramic -metal tetrodes for Apr distributed -amplifierserview 1ivaA1, C. E. Doner W. R. Weyant k1V

Distributed amplifiers have become an important part of electronic countermeasures systems because they prov de extremely broad bandwidth, high power. and high gain at frequencies up to approximately 400 MHz. In selectinetillrilloper amplifier device for these systems, such factors as bandwidth, power output, gain efficiency, environmental requirements. reliability,size. weight, and cost must be considered. This paper describes two RCA ceramic -metal beam podes designed specifically for high -power, broadband operation to 400 MHz indistrib- uted -amplifier services. Special design features of these tubes include an internal screen -gridbypasscapacitor,low -inductancegridleads.ruggedizationofth internal structure, and use of a monolithic heater -cathode assembly. Manufacturing techniques, process controls, and testing criteria are also discussed.

DISTRIBUTED AMPLIFICATION isa C E. Doner, Ldr. ProfessionsEngine3r. He is the holder of Leo Circuit Product Engineering patents cov3ring ttr.:e design techniques and f a> method of combining the out- Industrial Tube Division several others pendirg. puts of many amplifier tubes by Electronic Components substitution of input and output ca- Lancaster, Pa. W. R. Weyant received the BSEE from Pennsylvania State Uni- Regular Power Tubes pacitances of these tubes as shunt ele- vers ty and joined RCA Lancaster in1955 as a Industrial Tube Division ments in two artificial transmission Power tube applicationengineer. From 1959 to Electronic Components 1962,Mr. Doner worked as a design engineer Lancaster, Pa. lines (Fig. 1). This technique permits responsible for development of several power tubes received the BSEE in 1955 from Indiana Institute a large number of tubes to be operated including a UHF ceramic -metal tetrode for high - of Technology. He joined RCA in 1955 as an engi- in parallel over a very large bandwidth power distributed -amplifier service.In1962,Mr. neer on the specialized training program, following Doner was appointed Engineering Leader in charge which he spent nina years with the industrial de- which is independent of the number of of a design group responsible for design and design group in the receiving tube engineering ac- tubes in the amplifier. velopment of numerous power tubes including a tivity of the Electron Tube DivisioninHarrison. 500C -watt tetrode for UHF television service.In N.J., working on the design and development of Power tubes for use in high -power, 1963,he was appointed Engineering Leaderin new electron tubes for specialized industrial and charge of Regular Power Tube Application Engi- military application. He was also the responsible high -frequency distributed amplifiers neering. In 1966, Mr. Doner was appointed to his design engineer for industrial and military current must have high transconductance, high present position, responsible for the design and products.In 1964, Mr. Weyant was transferred to current capability, low lead induct- development of RF power generating devices in- regular power tube group, Lancaster Plant, where cluding strip -line and cavity amplifiers and oscil- he is presently serving in a similar capacity. Mr. ances, low input and output capaci- lators for CW, AM, FM, SSB and pulse applications Weyant has contributed to the quality improvement tances, low grid conductance, and a from HF through UHF. Mr. Doner is a Registered of industrial tubes and holds related patents. high degree of grid -to -plate isolation. These characteristics determine maximum operating frequency, bandwidth, efficiency, power output, power gain, and stability. Because these parameters are not all compatible, a tube designed forhigh -power distributed -amplifier service is optimized for the particular application.

Forced -air cooled tetrode The RCA 4624 forced -air cooled tetrode has been designed to meet the requirements of the distributed amplifier. This tetrode provides the highest practical power output with bandwidths as high as 200 MHz, and has an upper frequency capability of approximately 400 MHz. Many of the tube parameters which affect distributed - amplifier performance have been optimized in this tube.

Seated at left is W. R. Weyant; C. E. Doner is or the right. Final manuscript received October 8, 1968.

29 chassis. The shape of the flange permits close spacing between tubes which reduces series circuit inductance to as low a value as practical.. The grid is cylindrical in the active area and is supported by three small ceramic insulating washers in the cathode ground plane. Fig. 3 shows the two diametrically opposed grid leads which are used for input -circuit connections for operation in a low- pass circuit configuration. In this connection, the low grid -lead inductance is utilized as a circuit element appearing as part of the series line inductance. This type of grid lead also provides lower seal capacitance than an all - concentric terminal arrangement such as that normally used in UHF power - Fig. 1-Simplified low-pass circuit for a distributed amplifier. tube design. The third grid lead provides a connection for a shunt Grid -cathode assembly sintered nickel powder cataphoreti- inductor for operation in a bandpass Design of the 4624 began with the cally filled with emissive carbonates distributed -amplifier circuit. In this op- cathode and the control grid because This cathode produces high current eration the inductance of the lead is theinput capacitance, being much densities, has a long reserve life, and again a useful part of the circuit. higher than the output of power tubes, resists damage from arcing. The diam- The heater is of monolithic construc- determines the frequency capability of eter of the cylindrical heat dam at- tion and consists of an insulated tungs- the tube. For a bandpass distributed tached to the bottom of the active area ten helix connected to a molybdenum amplifier, the bandwith, f,-1 is de- of the cathode is smaller than the active heater rod at one end and to the heater termined by the following relationship: area to produce a minimum grid -to - cup at the other. The heater element cathode capacitance. A flat circular 1,- b = 1/7r(LC)"=1/1r(L,C,)" is held firmly in place by sintered nickel disc serving as an RF ground plane and powder filling the void between the where f, and f, are the upper and lower as a part of the vacuum envelope is heater cup and the cathode. This ar- cutoff frequencies, respectively; L, and attached to the bottom of this heat rangement provides an extremely Cp are the series inductance and shunt dam. This design provides anex- rugged heater. capacitance of the plate; and L, and tremely low cathode -lead inductance, C, are the series inductance and shunt which increases the upper fre- Screen -grid bypassing capacitance of the control grid. C, and quency capability and reduces input The screen grid C, represent the tube output and input conductance. is bypassed to the cathode inside the tube envelope to capacitance, respectively. The RF ground plane, or cathode minimize screen grid lead inductance. A minimum practical value of L, was flange, is constructed with four holes The screen grid consists of a cylinder estimated to be approximately 25 for direct mounting to the amplifier in the active area attached to a flat nanohenries. With f, at 400 MHz and chassis. An RF gasket provides a good washer -like support. This washer 1, at 200 MHz, the maximum permis- RF contact between the flange and the forms one electrode or plate of a radial sible value of C, is approximately 40 pF. After analysis of various cathode EXHAUST TUBULAT ION and control -grid configurations, it was

determined that a cylindrical cathode ANODE design having a minimum area of 3 SCREEN -GRID

cm', with a grid -to -cathode spacing of _CONT ROL - GRID 0.008 in. would provide high current CATHODE capability (steady-state plate current

of 350 mA in class -A service) and high SCREEN -GRID TO CATHODE transconductance(25,000 ,umhos at BYPASS 300 mA) and also meet the L, and C, CAPACITOR

requirements. NEATER -CATHODE TERMINAL

CONTROL -GRID Fig. 2 shows a cross section of the 4624 TERMINAL SCREEN -GRID tube (with the radiator removed). The TERMINAL cathode surface is the matrix unipo- NEATER TERMINAL Fig. 3-Two diametricallyopposed Fig. 2-Cross section of the 4624 tube with grid leads for input -circuit connections tential type consisting of a layer of radiator removed. in a low-pass circuit.

30 bypass capacitor in which a 0.003 -inch - top of the assembly to assure rugged- pating capability of 400 watts. With thick mica washer serves as the dielec- ization and alignment of the control recent requirements for 1000 -watt dis- tric. Another metal washer forms the grid with respect to the screen grid and sipation, however, air cooling becomes other electrode and attaches to the the cathode. The grid wires are formed impractical because the high air vol- cathode ground plane. A wire -mesh by electrical -discharge machining and ume needed requires high blower gasket, compressed during tubeas- are further refined by electro-polishing power. As a result, an intense -liquid - sembly, holds the capacitor parts in a which provides a ruggedized structure cooled version of the 4624 was de- direct contact. This bypass design pro- of precise alignment. signed to increase power output while duces an extremely low value of screen maintaining frequency characteristics grid lead inductance which reduces Plate design and compact size. Fig. 4 shows the undesirable feedback effects to a mini- Factors affecting plate design include single -pass, low-pressure cooler used mum. the plate -curve "knee" characteristics, for this tube, the RCA 4636. This A single screen -grid lead is brought DCplate voltage, and plate cooling and cooler operates with Coolanol-35, a through an insulator in the cathode output capacitance. The final design silicate -ester base dielectric fluid, at flange for application ofDCscreen -grid has an output capacitance of approxi- temperature from -50 to 135°C. The voltage. For low -frequency operation, mately 6.5pF. This relatively low coolant enters through theVginch where the 600-pF internal capacitor value enhances the high output -power lines and passes through a restriction does not provide sufficient bypassing, capability because output power is (for increased surface velocity at the additional capacitance may be added inversely proportional to output higher -thermal -density areas)in the externally to this lead. At frequencies capacitance. finned area. At an inlet temperature of 105°C and a dissipation of 750 watts, when this becomes necessary, the in- The five -piece plate assembly consists the required flow of Coolanol-35is ductive reactance of this lead is of the pinchoff tubulation, a heavy negligible. 0.75 gal/min at a pressure drop of 03 OFHCcopper cup for uniform heat dis- lb/in' (gage). OUTLET COOLANT LINE tribution, a Kovar stress -isolating ring, INLET COOLANT LINE a high -alumina (low-RF-loss) ceramic The 4636 is tested for a zero -bias cur- insulator, and a copper -clad Kovar rent of 1.4 A to assure satisfactory per- flange that serves as a connection to formance at a plate current of 600 mA the ground plane. This assembly is fab- in class -A, distributed -amplifier opera- ricated by means of a furnace braze tion. Typical operation of the 4636 in at 800°C. an 11 -tube bandpass distributed amplifier covering a frequency range of In the final assembly of the tube, the 107 to 187 MHz is shown in Table I. ground plane and the screen -grid by- The values shown are for the output pass capacitor are formed when the tube at an operating frequency of 141 OFHCcopper backup washers, the MHz and a plate -line impedance of gold -evaporated mica dielectric, and 250 ohms. the gold-plated wire -mesh gasket Fig. 4-Single-pass low-pressure cooler for Table Irepresents only one of many the RCA 4838. (which provides the proper compres- sive force and contact area) are placed pOssible modes of distributed -amplifier The control -grid -to -screen -grid spacing on the screen -grid flange. The plate operation. Although a broad discus- and the pitch of the grids were chosen assembly is then heliarc-welded to the sion of theoretical and practical tech- to provide a screen -grid -to -control -grid cathode flange. niques of distributed amplification is amplification factor of 12. This rela- beyond the scope of this paper, numer- tively high value permits class -AB, The tube is evacuated to approximately ous excellent papers have been written operation (zero grid current) at full 106Torr, and its plate temperature is on the subject." plate -current ratings with aDCscreen - raised to 450°C. The cathode is then grid voltage of only 400 volts. Zero - activated, and current is drawn to the References grid -current operation is essential to various tube elements to provide high 1. Ginzton,E.L.,Hewlett, W. R.,jasberg, prevent extremely high attenuation of temperatureforproperoutgassing. J. H.. and Noe, J. D., "Distributed Amplica- After this step is completed, the tube lion." Proc. of f.R.E. (Aug. 1948). the drive signal as it passes down the 2. Medds, S.K., An Experimental Distributed input line; low -screen -grid operating is "tipped off" (sealed by a metal clos- Power Amplifier, Naval Research Laboratory ure of the tubulation) and "aged" for Report 4985, Washington, D.C.,(Aug.20, voltage is important for increased am- 1957). plifier efficiency. stabilized operation. The tube is then 3. Grimm. A. C., Internal Correspondence (Feb. 26. 1957). Both the control grid and the screen dynamically tested at input frequencies of 30 and 600 MHz; shock -tested at Table I-Performance of RCA -4636 in grid consist of drawn cups of a heat - 11 -tube bandpass distributed amplifier tempered copper alloy. These elec- 30g, 11 milliseconds; vibration -tested at10g, 500 Hz; and life -testedat DC plate voltage 1800 V trodes are furnace brazed to the base oc screen -grid (grid -No. 2) voltage 400 V 600 MHz. occontrol -grid (grid -No. 1) voltage -60 V pins and the respective grid flanges Steady-state plate current 475 mA using a high -temperature silver -alloy Steady-state screen -grid current 10 mA Higher -power tube Steady-state control -grid current 0 mA solder; high -alumina spacers are used Peak RF drive voltage 55 V as insulators. A ceramic pin with anti - The 4624 is designed for forced -air Tube power output 365 W Useful power output (from 11 -tube leakage grooves is also brazed at the cooling and has a plate power -dissi- distributed amplifier) 2000 W

31 Klystron for the Stanford two-mile linear accelerator

A. C. Grimm F. G. Hammersand

The completion of the two-mile linear accelerator at the Stanford Linear Accelerator Center in the fall of 1967 provided physicists with the largest and most costly research tool yet built in the U.S. The accelerator center (Fig. 1) constructed at a cost of $114 millions provides the physicist with a 20 GeV electron beam and research facilities to probe the interior of atomic nuclei inhis quest to understand the fundamental nature cf matter. This paper describes the development of this linear accelerator with particular emphasis on RCA's rolein developing the high -power klystron for this application. Fig. 2-A 10 -ft section of accelerator wave - guide.

A. C. Grimm, Adm. Industrial Market Development Industrial Tube Division Electronic Components Lancaster, Pa. received the BSEE from the Polytechnic Institute of Brooklyn in 1940. Upon joining the RCA Tube Divi- sion in 1940, Mr. Grimm was assigned to the development of receiving and small power tubes. In this assignment from 1940 to 1951, he advanced to engineering leader. From 1951 to 1953, Mr. Grimm directed the advanced development effort on the RCA color kinescope. He was promoted to Product Manager for the Color Kinescope Product Group in 1954.In 1955, he was named Manager of Ad- vanced Development inthe Power Tube Group. In this capacity he was responsible, among other programs, for the development of the RCA 8568 klystron used to power the Stanford two-mile linear accelerator. From 1960 until he was appointed to his present position in 1967, he was manager of engineeringandoperationsmanager for RCA super power tubes. He has been issued six patents Fig. 1-The Stanford Linear Accelerator Center. Electrons begin their acceleration at the upper on electron tubes and associated devices, and he right and end their flight in one of the target areas of the "switch yard" in the lower left. haspublishedseveralpapers on the developmental aspects of vacuum tubes. He is a member of Tau Beta Pi and Eta Kappa Nu, and is a senior member of the Institute of Radio Engineers. He is a registered professional engineer In the State of Pennsylvania.

Fred O. Hammereend, Mgr. Super Power Device Engineering Industrial Tube Division Electronic Components Lancaster, Pa. received the BSEE and MSEE degrees from the University of Washingtonin1949 and 1956 respectively. Since joining RCA, Lancaster in 1951, he has been responsible for the design, development, and application of many gridded, gas, klystron, and magnetron tubes. This experiencein- cludes the design of electron guns for round and sheet electron beams, RF interaction circuits for gridded, klystrons, and magnetron tubes and tube and gun structures for the production and conservation of gas plasmas. Mr. Hammersand became a member ofthe A1230D klystron development team in December, 1960.InApril, 1966 he assumed group leader responsibility for the development of super power klystrons, magnetrons, and low frequency gridded tubes.InFebruary, 1967 he became manager of Super Power Devices De- sign Engineering. Mr. Hammersand is a registered Professional Engineer in the state of Pennsylvania; a member of the IEEE, Tau Beta Pi, and Zeta Mu Tau: and an associate member of Sigma Xl.

The two authors, A. C. Grimm (left) and F. G. Hammersand, are standing to either side of the RCA 8568 klystron.

32 THE CONCEPT for the two-mile lin- erator waveguide. Earth fill separates ear accelerator originated with a the concrete tunnel from the klystron group of faculty members at Stanford gallery and acts as a radiation shield University twelve years ago. Formal to protect workers in the gallery from proposals were made to the Atomic gamma radiation produced in the Energy Commission (AEC) on Sep- accelerator waveguide by stray elec- tember 15, 1961. After four years of trons. A portion of the gallery with an political and scientific debate, Con- RCA 8568 klystroninpositionis gress authorized the creation of a shown in Fig. 3. Fig. 3-A section of the Stanford Linear Ac- celerator klystron gallery using RCA -8568 national laboratory for the accelerator klystrons. to be called the Stanford Linear Ac- The RF energy from each klystron tube celerator Center (sLAc). is fed to the accelerator waveguide sec- 2856 -MHz pulse amplifier using a 2- tions through precision WR284 OFHC microperv, 250 -kV, 250-A electron Construction of the Center began in copper waveguide. The pressure level beam. Double-waveguide output was July, 1962. In May, 1966, the electron inside the entire waveguide system is specified at the time because there was beam was first activated. By January maintainedat approximately 1x10' some doubt that a single output win- 1967, a beam energy of 20.16 billion Torr. For minimum cost and mainte- dow could reliably handle the full out- electron -volts (20.16 GeV) and design nance, no vacuum windows other than put power. beam currents of 30 mA peak were those in the klystron tubes are used in The basic electrical characteristics as achieved. Beam currents of 45 mA the waveguide system. Waveguide vac- listed in the original tube specification peak have been reported recently. uum valves are used instead to close were as follows: off individual sections of the accelera- Operation of accelerator tor waveguide RF feed system when Operation A klystron tubes must be changed. Frequency (MHz) 2856 2856 In the SLAC system, a beam of elec- Peak power trons is accelerated through a two -mile - output (MW) 12 24 long waveguide comprised of more Klystron requirements Average power than 22,000 cavities brazed together Prior to Congressional authorization output (kW) 10.8 21.6 in 960 ten -foot -long sections (Fig. 2). RF pulse for the construction of the Center, the length (as) 2.5 2.5 After being fully accelerated, the elec- AEC had authorized limited funds for Beam tron beam can be switched to any of studies of the critical components for voltage (kV) 195 248 three experimental stations by large the system. The most important of Beam current (A) 172 247 electromagnets. Thus, one or more these components is the klystron tube Duty cycle 0.009 0.009 Gain (dB) 47 50 experiments can be conducted simul- used to power the accelerator. taneously by programmed beam Efficiency (%) 35 38 switching. The physicist's need for a highly stable Permanent -magnet focusing of the electron beam of low energy spread klystron electron beam was specified to The 20 -GeV energy level is obtained placed rigid requirements for stability reduce operating costs and to achieve by acceleration of the electron beam on the klystron tubes. Using AEC study a higher degree of reliability than pos- by 5000 -megawatt (MW) bursts of funds, engineers, and physicists at sible with electromagnet focusing. The 2856 -MHz power supplied by 240 klys- Stanford University designed and built use of permanent magnets eliminated tron tubes spaced at 40 -ft.intervals several prototype klystrons and devel- the need for 240 DC power supplies to along the accelerator. Each klystron oped tube performance specifications energize electromagnets. The savings provides a 21 -MW burst of power for more stringent than those required for obtained included not only the cost of 2.5 /As. The burst from each klystron military radar applications to assure the DC supplies, but also the cost of is accurately phased so that the RF satisfactory performance of the accel- electrical power consumption in ex- energy from the tube arrives at the erator. Tube cost, efficiency, and reli- cess of 250 kW that the electro- accelerator in the proper time sequence ability were also prime considerations magnets would otherwise consume and to accelerate the electrons as they pro- because of the large number of tubes the cost of an increased heat -exchanger gress down the waveguide. In the fu- used to power the machine. capacity of more than 250 kW needed ture, additional klystrons can be added In mid -1960, industry was requested to cool the electromagnets. Finally, to the machine to obtain an even by SLAC to complete the design of the the use of permanent magnets higher -energy beam. Provisions have prototype tube so that it could be re- eliminated troublesome water -coolant been made in the initial design of the produced and rebuilt at modest costs, connections that would otherwise be accelerator for a total of 960 klystron and to develop a permanent -magnet needed to connect electomagnets to tubes. focusing system for the tube. Both The klystrons, modulators, and power RCA and Sperry -Rand Corporation supplies are housed in a building called were successful bidders. Each was the klystron gallery. This building is awarded contracts to deliver six devel- located approximately 25 ft above a opmental tubes and to develop a per- concrete tunnel containing the accel- manent -magnet focusing system. The tube, as originally conceived by Final manuscript received October 10, 1968. Stanford, was a five -cavity, 24 -MW,

Fig.4-A half section of the RCA 8568 klystron fourth cavity showing 33 cavity loops and mode suppressors used to stabilize the tube. could be built only if theRFinteraction length of the tube was shortened. This change would reduce the length of the magnetic path and bring the volume of magnetic material required to produce the needed field within practical ALNICO B BODY OF PERMANENT bounds. Consequently,SLACbuilt a MAGNET satisfactory tube approximately 3 -in. shorter than the prototype.

PERMANENT -MAGNET MOUNTING FLANGE On the basis of this information, the

LEAD B -RADIATION SHIELD industry bid on production quantities UPPER CORONA SHIELD of both tubes and permanent magnets COLD -ROLLED STEEL CYLINDER in late 1962; again, RCA and Sperry - RCA -B568 BLYSTRON Rand were the successful bidders. ELECTRON GUN BOTTOM CORONA SHIELD 8 CATHODE -HEATER CONTACT Fig. 6-The RCA -8568 klystron installed in a Production klystron-RCA 8568 Fig. 5-Outline drawing of the RCA 8568 in- barrel -type permanent magnet. stalled in a barrel type permanent focusing The RCA 8568 klystron was designed were electromagnetically focused. magnet. A cut -away view is shown of the SLACspecifi- electron gun region. to conform to the revised The tubes also met all the required cations. For this design, the A1230D the coolant system. This arrangement mechanical specifications and had con- was modified to use one output wave - structional features that made it feasi- reduced maintenance problems and guide and window, and theRFinter- improved the system reliability. ble to rebuild them at low cost. action length was decreased by 3 -in. For additional reliability and ease of By the time the contract expired, satis- to make permanent -magnet focusing maintenance, it was specified that the factory operation to 16 MW had been practical. klystron tube have only two coolant achieved by use of a prototype perma- Although the above modifications connections, operate in a stable man- nent -magnet focusing system. Although seem modest, they changed the stabil- ner at any beam -voltage level between this performance was somewhat short ity characteristics of the tube drasti- 150 and 250 kV when operated at a of the goal of 24 MW, it was a record cally. Strong spurious oscillations in fixed focusing field, and be inter- achievement and demonstrated that, the frequency band between 5700 and changeable between focusing magnets, with further refinement, permanent - 6300 MHz and 5 -to -15 -MHz ripples on and that the output waveguide flanges magnet focusing was indeed feasible. theRFoutput pulse were chief among align automatically with the mating Encouraged by the results obtained by a host of instabilities observed in the flanges on the machine within a few modified A1230D. It took nearly eight RCA,SLACissued specifications in the thousandths of an inch. It was speci- months of careful analysis and patient fied further that the X-radiation shield- request for production tubes to power the system requiring a power output cold -test work for final stabilization of ing for each tube occupy the minimum the shortened design. During the of space and that the shielding reduce of 21 MW with permanent magnet focusing. Sufficient evidence had been course of this work, it was determined the X-radiation level to a value of 5 that the instabilities resulted from the obtainedatSLACon resonantring milliroentgens at 24 -in. from the tube following causes:RFfeedback through tests to suggest that a singleRFwin- axis to permit maintenance personnel the drift tubes; reflected electrons to work safely in the klystron gallery dow, coated with an extremely thin layer of titanium compounds, could drifting counter to the normal electron when the klystrons were operating at beam; and electron -beam instabilities maximum output. handle the full peak and average power from a tube. Therefore, for caused by transverse fields in the focus- Because the replacement of klystron minimum costs of both the accelera- ing magnets. tubes is a significant part of the operat- tor and the klystron tubes, production Solution of these instabilities included ing cost of the machine, a design goal tubes were specified to have only one judicious perturbation of the various of 2000 hrs (minimum) was estab- RFwindow. This change reduced the cavities, attenuation of the spurious lished for tube life.It was further number of vacuum valves and output resonances, improvement of the elec- specified that the tube be constructed flanges in the accelerator waveguide tron -gun optics, and development of in a manner that would permit it to be feed system by a factor of two and permanent magnet with an extremely rebuilt at a fraction of the original cost. saved construction costs. low transverse magnetic -field compo- Klystrondevelopment nent. Fig. 4 shows one half of the Experience gained as a result of the fourth cavity, and illustrates the system During 1961 and early 1962, RCA RCA work on the A1230D, together of cavity loops and mode suppressors redesigned theSLACprototype tube to with similar work done atSLACand devised to stabilize the tubes. Similar meet both the mechanical and electri- Sperry -Rand, indicated that it would mode suppressors are used in the other not be possible to construct a perma- cal performance specifications. Six four cavities. samples, designated RCA Dev. No. nent magnet of reasonable size and A1230D, were produced. These sam- cost to provide the requisite magnetic The barrel -type permanent focusing ples met the objective electrical per- field to focus a tube like the A1230D. magnet for the RCA 8568, shown formancespecificationswhenthey It was evident that a practical magnet mounted on a tube in Figs. 5 and 6,

34 weighs approximately 1300 lbs. About speed steel provides the pressure to 900 lbs of Alnico 8 are used to pro- compress anannealed,0.010 -inch - duce the field. A cold -rolled steel cylin- thick,OFHCcopper sleeve against the der is located beneath the bottom pole smoothly ground edge of the ceramic plate of the magnet barrel to shape window. This construction is econom- the axial magnetic focusing fieldin ical and eliminatesRFcurrent losses 150 in 200 2 BEAM VOLTAGE- kv the region of the electron gun and to that would otherwise occur in a brazed Fig. 6-The RCA -8568 klystron installed in a eliminate the bottom pole piece around metallized joint. acteristics. the klystron tube. TheRFoutput windows in the RCA - operating levels between 200 and 250 The axial magnetic -field shape for the 8568 are subject toa phenomenon kV. On the basis of the current data, permanent magnet was originally de- called single -surface multipactor. This which isstill incomplete because of rived from an electromagnet which phenomenon is manifested by localized insufficient tube failures, the projected had been used initially in evaluation window heating caused by bombard- life expectancy of the RCA -8568 is in of each tube. Thus, the early models ment of the ceramic by high-energy excess of 11,000 hrs. Many tubes have of the permanent magnets featured a electrons that become trapped in the operated for more than 10,000 hrs. ring of bar magnets inside the cold - RFfield near the window surface. The TheRFpower output, efficiency, and rolled steel cylinder to provide field phenomenon isself-sustaining when gain of a typical 8568 klystron with shaping which more closely simulated the secondary -electron emission char- permanent -magnet focusing are shown the electromagnetic field. After several acteristics of the window surfaces ex- in Fig. 7. Because of the fixed mag- magnets were built, it was found that ceed unity. If allowed to persist at the netic focusing conditions imposed by these bar magnets introduced non -sym- higherRFoperating levels, multipactor the permanent focusing magnet, the metrical radial magnetic fields which causes the tube to fail as a result of output -cavity design had to be com- degraded the klystron performance. window fracture, caused by overheat- promised to meet the dual power - When the bar magnets were removed, ing, or small vacuum leaks, caused by output specification; as a result, the the klystron tubes performed better electron drilling. operating efficiency of the tube typi- than they had in the electromagnet. Because both sides of the output win- cally peaks at 225 kilovolts. With early permanent magnets, non - dow operate in vacuum environments, Higher power -output levels than those symmetrical radial magnetic fields they are subject to the multipactor shown in Fig. 7 have been realized caused by transverse fields were a phenomenon. The problem has been with the RCA -8568 under experimen- major cause of poor klystron opera- eliminated by vapor -deposition of an talconditions. With normalpulse tion. Even with an axial magnetic -field RFtransparent coating of titanium lengths,RFpower outputs in excess distribution approaching the electro- compounds on each window surface of 25 and 30 MW have been obtained magnet field, it was impossible to to reduce the secondary -emission char- at RCA with beam voltages of 250 and achieve comparable performance with acteristics of the window surfaces. 275 kV, respectively. The highest a permanent magnet until the magni- Improved techniques are now being power output achieved experimentally tude of the transverse magnetic fields investigated as possible methods of im- with an RCA -8568 klystron was 105 was limited to less than 1% of the proving the coating stability. MW at a beam voltage of 425 kV and main axial field. Successful reduction a pulse width of 200 ns at MIT. The of the transverse fields by RCA and the Conclusions longest pulse operation was also magnet supplier, Crucible Steel, made This paper has discussed a few out- achieved at MIT when a tube was op- it possible for RCA to deliver the first standing design accomplishments of erated at a peak power output of 10 full -performance klystron to the Stan- the RCA -8568 klystron. Other accom- MW for 50 p.s. ford Accelerator Center by January, plishments in this work include the 1964, approximately a year before com- The RCA tube has received excellent design of a novel electron -gun con- customer acceptance. In the U.S., the petition. The uniformity of both the struction for easy repair, a coolant tubes and the magnets has made it RCA -8568isused atthe Stanford system to meet the Stanford specifi- Linear Accelerator Center, the Brook- practical to interchange tubes and mag- cations for two coolant connections, nets at the accelerator site. No adjust- haven NationalLaboratory, the X-radiation shielding to limit the X- Cambridge Electron Accelerator, the ment is needed on the magnets to radiation levels to 3 milliroentgens/ obtain full performance from any tube Cornell University Linear Accelerator, hour, and a unique suspension design operated in any magnet. and the Argonne National Laboratory. for mating the tube output flange to In Europe, the RCA -8568 is currently Output windows the accelerator. used at IKO in the Netherlands and A unique output waveguide window More than 152 RCA -8568 klystrons at the University of Ghent in Belgium. was developed for the A1230D klys- have been operated on the Stanford A variant of the tube for operation at tron' and was used without basic machine since its completion in 1966. 2998 MHz is being built for the changes on the RCA 8568 klystron. Although not all of the tubes have ORSAY accelerator in France. This window is unique in that the been operated atthe maximumRF Reference ceramic -to -metal vacuum jointis power level of 21 MW, statistical life - I. Teno, E., Hoffman, F. J., and Grimm. A. C test studies at Stanford indicate little "Dielectric -To -Metal Compression -Band achieved solely by means of pressure. Seals", A compression band made of high- difference between life of the tubes at

35 nical education and knowledge of this Environmental engineering growing field were considered essential for better understanding and guidance to the various product lines. These laboratory factors played a major role in the evolution of a centralized environmental J.M.FormanJ. B. Grosh facility supported by an environmental engineering staff. RCA in Lancaster is a major supplier of power tubes, display storage tubes vidicons, image orthicons, image converters, photomultipliers, and thermionic energy con- Current facilities verters for industrial,military, and space applications. This paper disc.isses the The procurement of a wide variety of facilities, responsibilities, lnd principles behind the RCA Lancaster Environmental specialized laboratory equipments for Engineering Laboratory that ,arts these products. Customer-orientec environ- simulation of stringent environmental mental tube applications are also described. field conditions led to the formation of a testing laboratory aimed at the eval- CURRENT SPACE AND MILITARY vacuum tubes in vibration and me- uation of electron tubes. Auxiliary op- SYSTEMSmust operate reliably chanical shock environments ac- erational equipments of flexible design in widely varied environments. A sin- counted for the vast majority of were engineered and built to accom- gle failure of a vital component may electronic system failures, began to modate numerous product -line types result in a costly loss of an airplane take corrective action. As a result, mili- under rated or above -rated electrical or a space vehicle. At the start of the tary contracts, starting in1955, retest conditions. Analytical electronic jet age and the space era a large num- quired that various product design instrumentation was also acquired or ber of component failures were attrib- groups at RCA Lancaster devote con- designed for monitoring tube perform- uted to the hostile environment. As a siderable effort towards the ruggediza- ance under environmental conditions. result,a new technology-environ- tion of electron tubes. Most of these Instruments such as spectrum analyz- mental engineering-was established. contracts stipulated that the tubes op- ers, accelerometers, charge amplifiers, One objective of this technology is to erate satisfactorily while being sub- and level recorders were found useful, subject system components totests jected to a variety of environmental not only for environmental tube anal- simulating the environmental condi- conditions. No single contract would ysis but also for acoustic noise meas- tions of their proposed mission; support the cost of all the environ- urements and for dynamic analysis of another objective is to analyze com- mental equipment needed to perform ponent failures and design various en- the required design and qualification vironmental protection methods and testing. Neither could the design ac- Natural, induced, and combined envi- devices. tivities send tubes, related electronic ronments can be simulated by these equipment, and personnel back and facilities for a variety of tube test pro- At present, many companies have grams: major investments in testing facilities forth from the plant to outside envi- ronmentaltestingfacilitieswithout )Design assurancetesting and eval- and technicalpersonnelspecifically prohibitive delays and costs. The need uation of partial assemblies or finished directed towards the field of environ- for the environmental facilities was tubes; mental engineering. Universities such 2) Up -grading of existing products; evident. as University of Pennsylvania, Massa- 3) Pre -production qualification; chusetts Institute of Technology, and From 1955 to 1958, RCA Lancaster 4)Reliability assessment and qualification; the University of Michigan are offer- acquired a number of environmental 5) Production sample testing; and ing environmental courses as part of equipments. More and more product 6) Qualification assurance testing. their undergraduate and graduate cur- lines gradually became involved in the ricula. In addition, a significant num- These environmental test facilities pro- use of these facilities. The custodians duce simulated environments such as ber of courses in shock, vibration, and of these equipments, receiving an in- vibration, shock, constant acceleration, acoustics are being offered by these creasing number of requests from the same schools in engineering summer temperature, humidity, and altitude. design activities, spent more time pro- Combined environments of tempera- conferences. The Institute of Environ- viding aid in calibration, maintenance, mental Sciences, a technical society ture -humidity, temperature -altitude, design of suitable holding fixtures, se- and temperature -vibration are also which has been in existence only ten lection of peripheral instrumentation, years, currently has a membership of available. All of these environmental and guidance in mechanical tube de- equipments and related instrumenta- over 2,500. sign for the dynamic environments of tion in the Lancaster Environmental shock and vibration. Evolution of the laboratory Engineering Laboratory are completely It soon became evident that environ- RCA owned. The military, keenly aware of the ac- mental problems were common to all celerating dependence of their opera- electron tube types. As a result, the Current responsibilities tions on electronics, and after having automatic transfer and integration of The environmental engineer provides determined that electrical relays and valuable design concepts from one consultation and technical guidance on product line to another were delegated tube design programs concerning en- Final manuscript received October 8, 1968 to a single engineering group. Tech- vironmental requirements. In this ca -

36 pacity, he offers engineering services sign. This development is followed by processing, automobile, steel, crime de- on customer proposal preparation, in- a detailed quantitative definition of tection, and educational institutions, cluding the development of environ- electrical and environmental limita- environmental engineering will play a mental test programs to meet planned tions of the tube. The customers are more vital part in the electronic in- ruggedization goals for a particular not impressed by the term ruggedness dustry. electron tube, according to customer unless it is backed by documented test Jules M. Forman, Adm. specifications. To accomplish this task, conditions and definitive tube ratings. Special Engineering Services he must have a familiarity with elec- Of the 450 tube types presently pro- Industrial Tube Division tron -tube principles and environmental Electronic Components duced at RCA Lancaster about 100 Lancaster, Pa. applications and a broad knowledge electron tube types have been evalu- ieceivei the ME from Stevens Institute of Tech- of environmental specifications and fa- ated by the environmental laboratory nology in1940. was a teaching fellow at Stevens cilities with associatedinstrumenta- from 1940 to 1941, and did graduate work there in the last decade. A tabulation of from 1940 to 1942 He joined the Equipment De- tion. He frequently assists design, some of the various kinds of field con- velopment Electrical Design Group at RCA Har- application,advanced development, ditions associated with these products risionin'941; and transferred to RCA Lancaster in1942,tothe Special Equipment Engineering and marketing personnel in the assess- is given in Table I. Group of the Life Test and Data Laboratory. Since ment and interpretation of customers' 1942. he has designed numerous electromechanical environmental demands. He is also re- Table I-Ruggedized Conditions electronic test sets, life test equipments, and has and Applications worked orthe specialapplicationof electronic sponsible for the design, drafting, fab- Ruggedized Field Conditions circuitry for small and large power, cathode ray, rication, and mechanical performance color, phoro and image. and display storage elec- Aircraft shock during landing tron tubes He was promoted to Leader of Spec al of the environmental fixtures required Constant acceleration during missile take -off in1951. Since 1956 to Lunarlanding(shortdurationhighimpact Equipment Engineering for tube analysis. include Special shock) 1962,hisactivity expanded to Shipboard shock impact from gun firing plat- Equipment and Environmental Engineering, having For developmental programs, the en- forms theadditionalresponsibilityforenvironmental vironmental engineer supported by Low -pressure -temperature conditionsofhigh engineeringi evalurr.ion and testing of all new and altitude aircraft improved tube ruggedization. From 1962 until his trained technicians performs all envi- Rocket engine random vibration forces recentappointment.Mr.Forman was Manager, ronmentaltests.Itislikewisehis Turbulent airflow around supersonic aircraft Environmental,SpecialEquipment 8Specifica- Ramjet screech tions Engineeringn the Electrical Measurements responsibility to evaluate the environ- Helicopter mechanically induced excitations due EnvironmentalEngineeringLaboratory.Mr. mental test results. From these anal- to rotary components and Shock impacts of camera tube housing by high Forman is a senior member of the IEEE, a member Sciences, a yses, he can furnish engineering speed focal plane shutter oftheInstituteofEnvironmental guidance to the tube designer in over- High impact shocks of naval ships that occur member of NSPE, and holds a Professional Elec- during a near miss of a torpedo which then ex- trical Engineering license in the state of Pennsyl- coming tube, deflection -yoke. mag- plodes and sets the urderstructure of the ship vania.Mr. Formanisalso aPastPresidentof netic -shield or other component into resonance. Lincoln Chapter, PSPE. encompassing Lancaster. Corona or breakdown created by low pressure Yorkand Adam Counties. ruggedization deficiencies. A further environment during operational conditions. responsibility of the engineer isthe Temperatures and humidities from the arctic to John B. Grosh, Ler. the jungle Environmental Engineering keeping of environmental facilities at Shock encountered in traversing rough terrain Industrial Tube D vision a minimum by guiding the product Shipping and handling environments Electronic Components Container drop tests Lancaster, Pa. lines toward standardization of test re- Railroad car or truck transportation received the BS in Physics from Franklin 8 Mar- quirements. Periodic calibration and Ruggedized Tube Applications shall College in June 1953. He then joined RCA as a Junior Engineer in Specification Engineering. regular maintenance of the equipments Army, Navy and Air Force airborne equipment After two yearsinthe service where he gained and instrumentationis sustained by Army half-track mobile equipment experience pertain ngtomissile telemetry and Tactical ballistic missiles these engineers and technicians to as- Oil well logging guidance systems, he returned to RCA in January sure the accuracy of their tests. The Sounding rockets 1956 and was assigned to the design of special Satellites electronic circuits and equipments in the Special fieldofenvironmentalengineering Weather EquipmentEngineeringLaboratory.Since13E7 covers many scientific disciplines: ther- Broadcast he has been active in environmental engineering modynamics, electronics, statics, dy- Reconnaissance related to electron tubes. He had been Acting Star -trackingforstellarnavigationcontrolof Engineering Leader oftheEnvironmentalEngi- namics, mathematical analysis, and the aircraft and missiles neering Laboratory since October 1962, and was Radar pulse compression named Leader in October 1968. He isa member environmental applications of mate- Airborne reconnaissance ofSigmaPiSigera,InstituteofEnvironmental Fire -control radar rials engineering. Finally, this engineer AmericanInstituteofPhysics,and Electronic countermeasures Sciences, is equally responsible for assessing the Night vision Society for the Acvancement of Management. stateoftheartof environmental Airborne range finding Television eye for: knowledge in the areas covering new Taxiing and landing of the supersonic transport equipment, new testing methods and Viewing exhaust of the Saturn rocket Observation of the earth and its resources procedures, and the theory behind the Observation of the launch of missiles test methods and procedures. Weather radar for commercial aircraft Underwater television for research and salvage Ruggedized tubes operations Ideally, the development of rugged- Future prospects ized tubes would involve concurrent In the coming decade, ruggedized and achievement of optimum environ- reliable electron devices will promote mental and electrical design objectives. the use of electronics in many commer- Unfortunately, these objectives con- cial fields. With the increasing use of flict; therefore a compromise is usually electronic products in such non -elec- made between the initial and final de- tronicindustries as chemical, food In the photo, Jules Forman (right) and John Grosh apply the finishing touches to an environmental test fixture. Design of a 915 -MHzpower triode for microwave cooking W. P. Bennett D. R. CarterI. E. Martin under mismatched load conditions as well as the application of computers to the design of transis- F. W. Peterson J. D. Stabley D. R. Trout tor amplifiers.

The microwave power system described in this paper uses a 915 -MHz power triode F. W. Peterson Advanced Development, Power Devices as the power source to provide an economical unit that has general utility. This sys- Industrial Tube Division tem provides a power output of approximately 900 W, and delivers more than 90% Electronic Components Lancaster, Pa. of full power to varying food loads. A large variety of foods have been prepared in received the BSEE from the University of Arizona the range. A 17 -pound refrigerated turkey has been roasted in 70 minutes. A com- in 1949 and has done graduate work toward the plete dinner for four, including chicken. baked potatoes, and a frozen vegetable, MS in Physics at Franklin and Marshall College. Mr. Peterson joined RCA-Lancaster as a Special- can be served in 25 minutes. Bacon is prepared to crisp condition in2 minutes, ized Engineering Trainee in 1949. Following com- and hot rolls and pastry require only seconds to heat. These results demonstrate pletion of his training program he was assigned to the Power Tube Development group. From 1950 to that a completely different, less time-consuming method of preparing food is at hand. 1954 he was responsible for the UHF circuitry used in evaluating the RCA -6161, 6181 and 6448 power W. P. Bennett, Mgr plicalion engineering. For the past seven years, he tubes. In 1954 he was assigned to the Power Tube Advanced Development, Power Devices has worked on special power tube problems and Design group as a Design Engineer where he Industrial Tube Division radio frequency circuit design, utilizing computer was directly concerned with the design of the Electronic Components techniques. In 1960 he was promoted to the posi- RCA -6816, 6884 and 7213.In1956 he was ap- Lancaster, Pa. tion of Product Development Engineer. In this ca- pointed Leader, Medium Power Design group. In received the BSEE from Michigan State University pacity he was assigned to work on the input circuit 1962 he was appointed Leader, Advanced Develop- in 1944 and has taken graduate courses in Physics designofa new developmental coaxitron Type ment, for Regular Power Tube Engineering. From and Chemistry at Franklin and Marshall College. A-2696 which he successfully completed. He has 1963 to 1967 he was responsible for the RF gen- Since joining RCA -Lancaster 23 years ago,Mr. since worked on other super power triode coaxi- erator development for cooking. Since 1967 he has Bennett has been engaged in the development of trons and tetrodes. In 1964 he was transferred to led a task force team involved in cost reduction power tubes and associated circuitry. He joined the Power Tube Advanced Development Activity of the RF cooking generator. Mr. Petersonis a member of Tau Beta Pi,Phi Kappa Phi, RCA in 1944 as a Product Development Engineer, and was assigned to the food electronics program, Pi Mu Epsilon and Sigma Pi Sigma. He holds one patent Industrial Tube Division at Lancaster. In 1956 he where he has been responsible for waveguide oven in the electronics field and has submitted several waspromotedtoManager,Super -Power Tube enclosure design and power triode oscillator cir- additional disclosures. He presented a paper on Development andin1958 was named Manager, cuitry. Mr. Carter is a member of IEEE, Eta Kappa the New Design Approach for Compact Kilowatt Super -Power Tube Design and Application Engi- Nu, and Sigma Pi Sigma. neering.He wasappointedManager,Regular UHF Tetrode at the 1958 IRE Wescon Convention. Power Tube Engineering in 1962 and assumed his I. E. Martin J. D. Stabley present post in 1963. Mr. Bennett was instrumental Advanced Development, Power Devices Advanced Development, Power Devices inthe design and development of of Industrial Tube Division a family Industrial Tube Division small -power,ultra -high -frequency Electronic Components triodetubes, Electronic Components RCA Types 5588, 6161, and their derivatives. Mr. Lancaster, Pa. Lancaster, Pa. received the BSEE from Tulane University in 1950 Bennett also directed an engineering group in the graduated from DeVry Institute of Technology in design and developmentofa of and the MS in Physics from Franklin and Marshall family ultra- 1950. Upon joining RCA in 1951 Mr. Stabley was high -frequency,high -power tetrodes capableof College in 1968. Mr. Martin has been responsible assigned to the Super Power activity under Lloyd continuous output of 25 kilowatts and peak power for the early stress analysis and design of ceramic - Garner.His assignmentsinclude work onthe output of 1megawatt. Mr. Bennett is a member of to -metal compression -typeseals and performed RCA 5831 usedin the Navy's Jim Greek trans- Tau Beta Pi and a senior member of the IEEE. considerable experimentationinconnection with mitter.He also worked on the development of He holds four patents and has several applications the development of these seals. He has studied the A-2332, a shielding grid Triode, used by the pending. He is the author of numerous technical heat -transfer processes in power tubes, both ana- A.E.C.inlinear accelerator applications at Berk- papers. lytically and experimentally. This work with both ley and the Lawrence Radiation Labs. In the late liquids and air,hasledto significant improve- D. R. Carter ments in the cooling of power tubes with liquids fifties Mr. Stabley participated in the design and applicationofseveralhighpower Advanced Development, Power Devices under extreme environmental conditions. Mr. Mar- Radar and Switch tubes that had applications in the BMEWS, Industrial Tube Division tin has participated in the design and development Nike Zeus and Tradex Radars. In the early sixties Electronic Components ofnumerous super -power tubes,particularlyin he was part of a team working under Navy contract Lancaster, Pa. connection with tube -processing, tube -circuiting, received the BSEE from Drexel Institute of Tech- and tube -testing. He has designed a number of to design, fabricate, and test the A-2696, a high nology in 1960 and the MS in Physics from Frank- circuits including a multi -megawatt cavity for UHF power broadband coaxitron for shipboard operation. Since 1964 he has been engaged in the food lin and Marshall Collegein1968. Upon joining pulse applications. More recently Mr. Martin has RCA in 1956 as a co-op student, Mr. Carter had been concerned with the generation of high power electronics program where his responsibilities are assignments in super power triode design, super through the use of transistors. He has also been in the area of enclosures, total system concepts, and customer contacts. power tetrode design, and regular power tube ap- concerned with modes of instability in transistors D. R. Trout Advanced Development, Power Devices Industrial Tube Division Electronic Components Lancaster, Pa. joined the RCA Electron Tube Division in 1954 as anUndergraduate cooperativestudent.He acquired thirty months experience during this period as an electrical equipment designer. Upon graduation from Drexel Institute of Technology in 1958 with the BSEE he was assigned to the Super Power Tube Equipment Development group and was instrumentalinthe development and operation of several high -power high -frequency test facilities. He iscurrently doing graduate work atFranklin and Marshall College for his MS In Physics. Mr. kr Trout is a member of Sigma Pi Sigma. The authors seated (left to right) are F. W. Peterson, I.E. Martin, W. P. Bennett, D. R. Carter, J. D. Stabley, and D. R. Trout. 38 THE USE of microwave power to 915 MHz. Thelonger -wavelength, grid lands to assure good electrical and heat food has many benefits as lower -frequency band seems to strike thermal contact. This feature greatly compared with conventional methods. a good balance for universal cooking. enhances the grid -dissipation limit to Its advantages include much faster The superior penetration of the long - the extent that the tube can withstand heating time, less food shrinkage, bet- wave energy and its general usefulness operation even in the completely un- termoistureretention, and several in cooking and heating both large and loaded condition. The cathode area of significantconveniencefactors.Be- small items are desirable features. In 10 cm' is divided into 36 equal sections cause microwaves primarily heat the addition, because a 915 -MHz power - to keep the span of the 0.0015 -in. grid food, not the container which holds it, generating device can be designed for wire very short. dishes may become warm but not hot operation at low voltages, the range Fig. 3 shows the component parts of enough to burn or to accumulate power supply can be a simple, light- the grid -block, cathode -block assem- "baked" foods on the surface. As a weight, inexpensive,transformerless bly. The gear -like grid block is manu- result, food items may often be pre- unit that uses low -voltage solid-state factured economically by the use of pared, stored, heated, and served in devices. For these reasons, 915 MHz cold -extrusion forming techniques the same dish, and the smaller number was selected as the frequency band for which repeatedly produce parts to the of dishes will be easier to clean. In ad- the design described. close tolerances required. The one- dition, dishes and containers made of piece extruded grid block contains 36 paper and plastic may be used to great Microwave generator lands on which the grid wire is wound advantage. The heart of a microwave heating sys- and 36 slots which serve as the bearing In spite of these advantages, the num- temisthe device employed as the surface for one end of the cathode ber of microwave ranges presently sold microwave generator. A power triode strands. An identical set of 36 slots for for use in the home is an almost in- was selected because of its potential cathode seating are formed on the significant percentage of the total low cost, small size, and light weight, smaller cathode block. The grid and range market. To increase this percent- and because of the proven production cathode blocks are guided on a com- age, suppliers of microwave sources technology already established for this mon axis by a ceramic rod through the must give greater consideration to two type of device. The 915 -MHz triode center of each block. requirements: power generator has the additional The total matrix -oxide cathode area of advantage of being able to tolerate 1) The economics of the microwave 10 cm' supplies a steady-state current power unit must be compatible with both unusual loads (including metallic of 0.5 A/cm', as required for low -volt- the values of microwave heating as utensils) and the `no-load", or empty, age (600 V) operation. This matrix - judged by the consumer. oven condition. Fig. 1 shows a photo- oxide cathode has been used exten- 2) The general utility and ease of use graph of the power triode; Fig. 2 illus- sively for years in many RCA power of the range unit without burdensome trates its construction. do's and don'ts must be consistent with tubes. In the 915 -MHz triode, directly the user price for the unit. The general design objectives for the heated strands are used to minimize microwave power triode were as fol- cost and also to achieve fast warmup. The strands are directly connected to Operating frequency lows: microwave power output of 1 kilowatt at 915 MHz, low -voltage the grid block and cathode block in The best frequency for heating food the tube. Direct connection is made is one which is high enough for the (600 -volt) operation, high efficiency, low cost, fast warmup, forced -air cool- possible by inductive "moats" in the wave energy to be converted to heat grid block. This design greatly simpli- energy by molecular motion in passing ing, service life of 15 years, light weight, and small size. In addition, the fies tube construction because only two through the food, yet low enough to insulating members are required in the penetrate thoroughly into large items tube would have to deliver essentially full power to a wide range of load im- vacuum envelope (the minimum re- such as roasts or fowls. The infrared quired even for a diode) , and the need waves used in conventional ovens have pedances representing the nearly in- finite variety of food loads anticipated. for a DC bias supply is eliminated. The very short wavelengths; they do not onlyelectricalservicesdirectlyre- penetrate the food but are converted A triode oscillator design was selected quired are a filament heating supply to heat energy in a very thin layer at to meet these objectives. Unique input and the plate voltage supply. These the surface. As a result, the inner por- and feedback circuits were incorpo- features contribute substantially to the tions of the food can be heated only by rated into the design so that the phase low-cost objective of the system. a slow process of thermal conduction. of the grid -cathode drive voltage could The anode for the triodeis also a Extremely long wavelengths, on the be properly fixed and the magnitude other hand, pass through food with es- cold -extruded copper cylinder which could be adjusted to optimize the effi- is formed with a thin -walled strain - sentially no heating effect. Either too ciency of the oscillator. short or too long a wavelength results isolation section on either end. The in inefficient heating. The tube design features a massive radiator consists of three concentric copper grid block, wound with copper - rings of corrugated copper strip brazed Two frequency bands assigned by the alloy grid wires, to maximize heat to the anode cylinder. In addition to FCC are presently being used for transfer to external parts that can be providing the thermal heat path to the microwave heating: 2450 MHz and maintained at low operating tempera- air stream, the fins of the radiator form tures by forced -air cooling. The grid part of the microwave circuit in the Final manuscript received October 8, 1968 wires are mechanically peened into the generator. Fig. 4 shows the mount as -

39 Fig. 3-Component parts of the grid -block, Fig. 4-Mount assembly ready for insertion cathode -block assembly. into the anode -radiator assembly. by the tube itself; the outer conductor which contains the output transmis- is a 5 -in. -diameter, 7 -in. -long cylinder sion line, forms a complete shield that made of two identical, die-cast, alumi- prevents the leakage of microwave num half -shells that provide rugged, energy. Because itis electrically iso- lightweight construction. Attached to lated from all of the tube electrodes, it Fig. 1-Completed triode as an enclosed each end of the tube are end -cup can be connected to the system ground. unit. blockers (metal cylinders covered with The generator weighs approximately sembly ready for insertion into the an insulating sheet) which mate to the 91/2 lbs., and occupies a small volume anode -radiator assembly. terminals of the tube, provide a micro- consisting of a 5 -in. -diameter cylinder Processing of the tube during exhaust wave short for the end of the coaxial 7 in. long, and a 51/2 -in. length of 15/8 - is conventional except that grid bom- line, and form aDCvoltage block on in. -diameter coaxial transmission line. bardment cannot be accomplished by their outer diameter. The bottoms of application of positive grid bias be- the end -cup blockers consist of 36 Triode microwave circuit cause the cathodes are internally con- radial fins which permit the free pas- The microwave circuit for the tube nected to the grid. Therefore, the grid sage of cooling air through the entire can be described electrically as a quasi- is bombarded under microwave oper- coaxial cavity. Colpitts circuit. For optimum perform- ating conditions. Consequently, a get- Air is directed into the cavity from ance of the generator, the microwave ter ion pump is connected to the tube the grid -block end, flows through the circuit must control the magnitude and until after a short period of microwave anode radiator, and exists past the phase of the feedback signal and the oscillator operation in a cavity similar cathode -block end. This direction of air system frequency. In the lumped -cir- to the one used for final generator flow provides cold air on the grid cuit model shown in Fig. 7, the values assembly. block which compensates for grid dis- of C L and the load Resistor Kr, are The microwave cavity for the oscilla- sipation and prevents temperature un- controlled by the circuit external to tor consists of a coaxial cavity on balance between the grid block and the vacuum envelope. The grid - either end of the tube with the anode the cathode block. cathode circuit, built entirely within the vacuum envelope, is shown sche- radiator in the middle. The center con- Adjustment of the electrical line ductor of the coaxial cavities is formed matically as C L and R. The capaci- length, established by the position of tor C2 represents the active portion of the end -cup blocker on the grid -block the grid -cathode electrode capacitance. end of the tube, determines the oper- EXHAUST The inductor Lg represents a short TUBULATION ating frequency of the oscillator. Sim- length of transmission line terminated FILAMENT ilarly, positioning of the end -cup in a short circuit where the cathode is LEAD STUD blocker on the cathode -block end of GRID BLOCK directly connected to the grid. The re- the tube establishes the magnitude of sistor R, represents input losses caused the feedback voltage and thereby de- R F CIRCUIT by grid absorption and energy losses CONTACT SURFACE OUTPUT termines the operating level of plate CERAMIC resulting from acceleration of the beam current. DIRECTLY HEATED B. LEAD electrons into the grid -to -plate region. CATHODE ONNECTOR Microwave energy is capacitively coupled out of the cavity by means of a1541 -in. -diameter, 50 -ohm transmis-

ANODE -RADIATOR CERAMIC sion line that enters the cavity perpen- ALIGNMENT PIN OUTPUT dicular to the axis of the tube in the CERAMIC RF CIRCUIT CONTACT SURFACE region of the radiator. Fig. 5 shows the EXTERNAL TENSIONING. - triode, with end -cup blockers and fila- SPRINGS -CATHODE -BLOCK ment leads attached, positioned in the lower "half -shell". Fig. 6 shows a com- FILAMENT LEAD STUD plete generator assembly, with the upper "half -shell" securely bolted to EXHAUST Fig. 5-The triode with the end -cup blockers TUBULATION the lower "half -shell". This outer shell, and filament leads attached.

14) Fig. 2-Cross-section of the power triode. IF v pit TRANSIT ANGLE Vgp Ip

Fig. 8-Oscillator phase diagram. the oven forms standing -wave patterns Fig. 7-Lumped-circuit model of the micro- which are determined by the boundary wave circuit. conditions: the oven dimensions, and between the computer -predicted data the size, shape, and characteristics of and actual experimental data. the food load to be heated. Because Fig. 6-A completed generator assembly. of these standing -wave patterns, the Automatic power regulator effective heating voltage varies and In a microwave triode oscillator, elec- Adjustment of a coupling capacitor in some portions of the food are heated tron transit time affects the boundary series with the output transmission more rapidly than others. To avoid conditions that the circuit must satisfy. line permits the power output of the this non -uniform heating, it is common For optimum power generation, the triode oscillator to be held at an opti- practice to move or change the voltage fundamental plate current I. of the mum value even though the load im- standing -wave pattern with relation to tube must be in phase with the funda- pedance varies widely. Such control the food, and thus to equalize the heat- mental grid -to -plate voltage V, of the isfeasible because maintenance of ing effect throughout the item being tube. The power generated PO is either constant steady-state plate cur- heated. given by rent or constantRFvoltage in the oscillator cavity produces optimum There are always a number of differ- P, = IVI 11,1 cos x power output in the load. As the load ent standing -wave patterns which are likely to occur, depending on the exact [The choice of V, instead of V com- is varied, therefore, the proper capacitor setting for optimum power can be frequency of the power source and pensates for the normal 180° phase the perturbation of the electric field in reversal at the output; therefore, the determined by a control circuit such as that shown in Fig. 9, which senses the oven by the food load and other negative sign in the expression drops oven parts. These patterns (or modes, out.] a change in either of these parameters and adjusts the coupling capacitor to as they are often called) can be excited where x is the angle between V and minimize this change. as a result of moving the food, pulling I, (ems values are used). At 915 MHz, the frequency, or rotating a metallic the plate current I. lags the grid -to - The effectiveness of this controlis stirrer inside the oven. The greater the cathode voltage in time because of the illustrated by the performance diagram number of modes or different field pat- electron transit time in the grid -to - shown in Fig.10. This diagram, terns in the oven, the greater is the cathode region. Therefore, the grid - which is essentially a Rieke diagram diversity of heating patterns and the cathode circuit is tuned precisely the for a power -regulated oscillator, shows better the likelihood that the average right amount below the desired oper- a set of power and frequency coordi- heating effect throughout the food will ating frequency so that the grid -to - nates superimposed on an impedance be more uniform. A greater number of cathode voltage leads the grid -to -plate plane (Smith Chart). It can be seen modes also reduces the range of im- voltage by a phase angle equal to the that 90% of full power can be pro- pedances presented to the power grid -to -cathode transit angle. This de- duced with standing -wave ratios as source by different food loads. high as 10 to 1; higher standing -wave sign compensates fortheinherent Because the number of standing -wave transit -time lag, and thus accomplishes ratios are seldom encountered, even in an empty oven. patterns is relatively low in the 915 - the desired objective of having I, in MHz band, a method was developed phase with V. Tuning of the grid - Another feature of this control is that for moving the food in relation to the cathode circuit to 825 MHz optimizes regulation can be applied to several field patterns to provide uniform heat- the power generated at 915 MHz, as discrete power levels. In most cases, ing. In this method, proposed by shown by the phase diagram in Fig. 8. two power levels are adequate and a Prucha,' the food load is slowly ro- A test procedure was established to switching arrangement is used to select tated in the oven on a turntable or verify the transit -time correction. The the desired level. A lower regulated "lazy susan". Either a dielectric mate- passive resonant frequency of the en- power level is achieved by a change rial is used for the turntable, or the tire circuit is the same as the oscillat- of the reference level of the sensing food is supported on an insulating plat- ing frequency only if the transit -time signal to select a lower current- or form 2 to 3 in. above a metal turntable. voltage -level reference. compensation is correct. The amount The voltage vectors of the field pattern of any correction necessary can be ob- Oven enclosures excited in the oven enclosure should tained from the difference between One problem in the design of micro- be predominantly in the horizontal these two frequencies and the loaded wave ranges is the excitation of the direction. This excitation patternis Go of the circuit. proper voltage patterns in the oven. particularly beneficial in coupling The dimensions of the circuit were cal- Any microwave oven enclosure is power into thin food loads such as culated on a computer to avoid lengthy basically a waveguide resonator; there- bacon. A simple and effective way of computations. Good correlation exists fore, the microwave energy fed into exciting such field patterns at 915 MHz

41 MERCURY PLUNGER RELAYS

RCA iNI206A AUTO LOAD system for the future when the eco- CONTROL nomics of the solid-state components 236V 4,60 H2 OVEN become more favorable. The tube and cavity require an air ANTENNA GND -N flow of 100 ft'/min and have a pres- RCA INI206A 10A TOG BLOWER sure drop of 0.35 in. of water at this CURRENT CONTROL flow. LAMP PHOTOCELL Ii5 v.--o-T7 NOTE AT ZERO PLATE CURRENT FILAMENT Fig. 9 also shows a rather unique cir- INPUT IS H30 A AT _ 07V cuit which maintains the filament heat- TIMER ing input at the optimum operating value for long life. This circuit regu- lates the filament power as a function of the anode current by means of an adjustable resistor in series with a Fig. 9- System circuit diagram. lamp. Higher -than -normal heater power is provided to decrease starting time, and then, as cathode current is

WAVE - - WAVE drawn, the heater power is automati- -LENGTH 0 LENGTH is to feed the power to the back of the 02 02 oven by means of a coaxial transmis- cally reduced to the minimum level re- TOWARD quired to maintain microwave power TOWARD sion line, and to excite the oven by LOAD GENERATOR output.(Thiscircuitissomewhat 05 05 means of an antenna probe which is simply an extension of the coaxial -line sophisticated and, for reasons of econ- center conductor, as shown in Fig. 11. omy, may be replaced by a simple relay circuit which drops the filament power -9MHE i -3MHz Power supply and back to a predetermined value once auxiliary equipment the tube has started to operate.) The advantages of 915 -MHz micro- Systemperformance wave energy and the potential econ- ALL LOADS WITHIN Fig. 12 shows a typical set of compo- THIS BOUNDARY 2 2 omy of the small, lightweight power PRODUCE 90- nents for the microwave heating sys- ,00 PER -CENT REF PLANE oscillator can be realized only if the POWER 3 remaining components of the system tem,including power -supplyparts, 5 filament transformer, blower, and 10 0 10 R f GENERATOR are also light in weight, reliable, and low in cost. If conventional 236 -volt triode power generator. Typical sys- Fig. 10-Performance diagram for the power - tem and tube operating conditions for regulated oscillator. household voltageisutilizedina standard doubler circuit, approxi- operation from a conventional 118/ 236 -volt, 60 -Hz, 3 -wire power line are RF PROBE mately 600 volts ofDCare made available to the tube anode. Such a circuit as follows: permits a substantial reduction in ACline current 17 to 19 A weight and cost through the elimina- Power factor 0.7 Heater current: tion of a power transformer. for starting 180 A at 0.7 V The power supply and auxiliary cir- for operation 150 A at 0.6 V DCtube anode voltage 625 V cuits shown in Fig. 9 have been life- Dc anode current 3.8 A tested under on -off cycled conditions Operating frequency 915 MHz without component failures for periods Starting time (approx.)20 s in excess of 3000 hrs. This life -test Generator air -flow

COAXIAL TRANSMISSION LINE period corresponds to an estimated requirements at TO GENERATOR standard conditions Fig. 11-Sketch of the microwave range. elapsed time of 30 years in a typical (at0.35 -inch water kitcheninstallation where usageis pressure) 100 ft'/min. about 100 hrs/year. Power outputto matched load 950 W An alternate power supply was de- Power delivered to signed in which thyristors served the water load in typical dual functions of switching and recti- oven enclosure 750 W fying, and thus replaced the mercury Total system weight for the generator, plunger relays and the diodes. With blower, filament transformer, and this solid-state system, a firing circuit power -supply componentsisabout was used to provide a "soft" start for 30 lbs. the system by charging the capacitors over many cycles of line voltage rather References Fig. 12-Typical set of system components, than during one cycle (as with relays). I. Prucha, H. V., "A New Look at the Electronic including power supply parts, filament trans- Oven," 16th Annual Appliance Technical Con- former, blower, and triode power generator. This alternate design may be the best ference, Columbus, Ohio (May 18-19, 1965).

42 RCA Lancaster 25 years of engineering excellence

At RCA's Lancaster plant-at the heart of Pennsylvania Dutch country- they certainly "know what g000d is." Lancaster engineers turn with just pride to their 25 -year record of firsts in power, conversion. and color -tube engineering. Some of these firsts are given in the chronology below.

The Lancaster plant (shown at right) has grown from a 350.000 -square -foot manufacturing facility in 1942 to a 1.400.000 -square -foot complex employing over 5200 in applied research. design, development, and testing. as well as production. Their products vary from TV color picture tubes and RF cooking tubes used in home appliances, to giant high -power klystrons and magnetrons for government and industry, and to sensitive image orthicons. vidicons. and isocons for the entertainment industry and the space program.

Cathode-ray and picture tubes

1943 Cathoderay tube engineering department transferred from Harrison to Lancaster. At that time,it included all eng'neering on cathode-ray tubes such as radar, oscilloscope, flying spot, and projection types, as well as all pick-up tubes, such as iconoscopes, orthicons, and image tubes. 1949 Tri-color picture tube engineering began in Lancaster. 1950 RCA demonstrated first tri-color tube and a compatible all -electronic color television system to the FCC. Pick-up tubes set up as a separate department. 1951 First of several RCA symposia for presenting detailed technical and engineering information on its color tube to competing tube and set manufacturers. 1953 15GP22 flatscreen color picture tube put into production. Development of NTSC color system completed; adopted by the FCC. 1953 to 1954 Transfer of black -and -white picture tube design to the Marion, Ind. facility. 1954 21AXP22 metal color picture tube announced and demonstrated: the first tube to use lighthouse lens correction. 1955 Color tube applications set up as a department separate from color tube design. 1957 2ICY22 introduced-the first color tube in all -glass bulb; it used "degrouping" lighthouse lens and graded apertures in steel masks. 1960 RCA introduced all -sulfide phosphors in color tubes, producing 50% increase in brightness (21FBP22). 1961 21E11'22 introduced-with bonded safety -glass. 1964 25AP22A introduced-first RCA 90' rectangular color picture tube. 1965 Introduction of Vanadate red. 1965 to presentImprovements in ed-emitting phosphors (Vanadates and Oxysulfides); introduction of Einzel-lens gun. 1966 Development and introduction of "Perma-Chrome" tempera -tube compensation system. 1966 Einzel lens introduced in the 15 -in portable color set. 1967 Development and introduction of Oxysulfide red-first "unity current ratio" color tube.

Super power devices

1937 Super power tube research and development activity was started by L. P. Garner and Dr. P. T. Smith under Dr. V. K. Zworykin. 1943 Advanced Development Group established in Lancaster under L. P. Garner. 1947 The first commercial 1MW CW beam power triode, the RCA -5831. was marketed; the U.S. Navy bought the first tubes for their Jim Creek transmitterin Washington State. 1950 to 1955 Development of the RCA -6949. 500kW CW shielded grid beam triode for high -power communications. radar, and particle accelerator applications. 1953 The RCA -6448. the first UHF -TV high power beam tetrode tube, was announced. 1955 to 1958 Development of the RCA -6952, the first high power pulsed UHF radar beam power tetrode utilizing matrix oxide cathodes based on the electronic structures conceived for the RCA -6448. 1958 to 1961 The RCA Developmental Type A2346, the doubled -ended UHF triode now commercially available as the RCA -2054 and 7835, was developed for the Air Force. This structure provides the highest average power available in the world from gridded tubes suitable for UHF operation. 1961 to 1964 Development of the first coaxrtron. integral -circuit high -power triode, under Air Force contract. Three Developmental Type A15038 Coaxitrons capable of 5MW peak power from 400 to 450 MHz were delivered. 1961 to 1964 The RCA -8568 30MW pulsed klystron for particle accelerator applications. 1964 to 1968 The development of heat pipes-devices for the efficient transfer of heat. 1966 Development of RCA -8684 25kW CW magnetron for industrial heating applications. 1966 Development of 100mW and 1W CW Argon lasers; a number of units have been produced and sold. 1968 Development o. aI kW CW klystron for operation between 4500 to 5000 MHz.

Regular power tubes

1952 Development of the 6146 tube family for VHF mobile communications. The 6166 tubes were developed for VHF -TV transmitters. The 6181 tubes were developed for UHF -TV transmitters. 1956 The 6816 tubes were the first of the cermalox family that were developed. 1960 First conduction -cooled cermalox tubes developed. 1961 Ceramic -metal 8072 developed for UHF mobile communications.

Conversion tubes

1962 Photocell facility moved to Mountaintop. Penna. 1965 Convex Expansion Program which included a Laminar Flow Clean Room (largest and cleanest in the world). 1966 Introduced long -life target image orthicon. increasing tube life four -fold. Great surge in image tube business because of night vision devices, increasing four -fold since 1955. 1967 TV guided missile business created an expanded vidicon tube market. Bi-Alkali photocathode was introduced to image orthicon and photomultiplier product Ines.It vastly increased the general performance of these related types. 43 The heat pipe,an unusual thermal device R. A. Freggens R. C. Turner

The transportation of unwanted heat to a suitable heat "sink" has proven to be a difficult design problem. Now. a unique device known as a "heat pipe"' can transport large quantities of thermal energy to remote heat sinks with a temperature difference of only a few degrees. As an example of this unique ability. 11.000 watts of thermal energy have been transported through a one -inch diameter heat pipe over 27 inches long, with a temperature loss so small that it was difficult to measure. By comparison. a copper bar nine feet in diameter and weighing about 40 tons would be rNwired to produce the same result. This paper describes the operation and design of feat pipes. and summarizes RCA's rolein the development of these unusual thermal devices.

R. C. Turner R. A. Freggens THEHEAT PIPE is basically a sealed Heart Transfer Devices Engineering Heat Transfer Devices Engineering two-phase system in which a fluid is Electronic Components Electronic Components Lancaster, Pa. Lancaster, Pa. continuously evaporating and condens- joined DEP's Missile and Surface Radar Division received the BS in Chemistry from Upsala College ing. The containment vessel shown in in 1958. He received the BSME from the Pennsyl- in 1951, and joined the Westinghouse Electron Tube Fig.1, is evacuated of all unwanted vania State University in 1961. and was reassigned Division as an engineer assigned to power tube to DEP's Astro-Electronics Division as a member of development, responsible for the development of gasesand sealed. A wick structure the Thermal Design Group. In this position he was power tubes for audio, RF and high -voltage switch lines the inside wall from the evapora- involved in the thermal design and environmental tube applications. His contributions included the testing of many of the spacecraft developed by design and improvement in electronic geometries tor end to the condenser end. Heat AED. including Ranger, TIROS, Relay and Nimbus. and processing of high -power griddect tubes, with supplied to the evaporator causes va- In 1965. Mr. Turner was transferred to EC where he particular emphasis on the interaction between porization of the fluid stored in the became a member of the Thermionic Converter elements within the devices. W. Freggens joined Engineering Group of the Direct Energy Conversion RCA in 1966. Since then he has been responsible wick. The heat is then transported to Department. Since then, he has been responsible for the development of high tempe'ature thermionic the condenser end in the form of vapor for the design and development of novel, fossil - energy converters, employing heat pipes to transfer flow, where it is recovered by conden- fuel -fired, thermionic power systems for the Army, the primary thermal -energy from a nuclear heat and a 500 watt, thermionic. radioisotope fueled source to a thermionic converter. He has been in- sation of vapor back into the wick. space power system for the Air Force. He is cur- volvedinspecial studies of molybdenum -lithium The cycle is completed when the con- rently responsible for the development of a 100 - heat pipes for thermionic systems, under current densate is returned to the evaporator heat -pipe Ranklin Space Power System Radiator for in-house contracts. He has developed special low - the Air Force.Mr. Turner has authored orco- temperature heat pipes employing aluminum and by capillary action through the wick authored several technical papers and has a patent copper as the envelope material. He has also in- structure. pending for a highly accurate thermocouple device. vestigated the characteristics of these units to ob- He was awarded a certificate of commendation by taincompatible combinations ofenvelopes and The purpose of the working fluid NASA/JPL for,"significant contributions tothe working fluids for achieving reliable lorg-life oper- within the heat pipe is to absorb the success of the RANGER 7 mission which secured ation.In addition, he has developed radial heat heat energy injected into the evaporator the first high resolution photographs of the surface pipes for use in specialized applications. Mr. Freg- of the Moon." and received one of the 1967 Science gens is a member of the IEEE, has been awarded section. This is accomplished by va- awards fromtheInventionsandContributions onepatent and hasthreepatentapplications porization of the working fluid, a proc- Board of NASA for "significant value to the ad- Pending, vancement of the NASA Space Program." cess that requires a large amount of heat energy, i.e., the latent heat of vaporization. (There is no increase in fluid temperature associated with this heat energy.) The vapor is transported to the condenser section by a slight pressure difference created between the ends of the pipe. When the vapor arrives at the condenser section, it en- counters a slightly lower temperature than that of the evaporator. Conse- quently, the vapor condenses, releasing the thermal energy that had been stored as heat of vaporization. For all practical engineering purposes, the heat pipe can be considered essentially isothermal along its length.

Final manuscript received October 17, 1968.

44 In the photo, Mr. Freggens (seated) and Mr. Turner are examining various heat -pipes. EVAPORATOR SECTION CONDENSER SECTION rWICK 4 / / / T// /- /// / Table I. RCA heat pipe developments Fluids -7- Methanol Ettylene Inorganic salts _ Acetone glycol Indium Water Mercury Lithitin Dowtherm Cesium Bismuth Napthalene Po:assium Lead

Sodium I If f fit I Vessels and HEAT INPUT HEAT OUTPUT wicks LIQUID VAPOR Glass Stainless steel Hastdloy X Fig. 1-Cutaway drawing of a typical cylindrical heat pipe. Copper Aluminum Alunina Nickel M3lybdenum (TZM) Geometries Cylindrical Radial 611(01, Double diameterSerpentine Double ended Re-entrant Star* Flexi3le Life continuing and undegraded Lithium-TZM 10,400 hors Sodlum-Hastelloy X 17,000hours Potassium -Nickel 15,000hours Mercury -Stainless steel 10,000hairs Water -Copper 5,000hairs Water -Stainless steel 4,000hairs

This nearly isothermal propertyis maintained even at high heat transfer rates and is responsible for the high Fig. 2-"Serpentine" heat -pipe. effective thermal conductance. Heat pipes have also been fabricated to weight and volume is an obvious de- transfer large quantities of heat while sign objective. The heat -pipe panel RCA developments withstanding high voltage potentials shown in Fig. 5 was used to provide Since RCA began research on heat applied between the evaporator and a high heat -rejection rate from a Rank- pipes in 1963, nearly 800 heat pipes condenser ends. One such pipeis ine Space -Power System. The radiator of various sizes, temperatures, fluids, shown in Fig. 3 The containment ves- subsystem incorporates 100 stainless - materials and power -handling capabil- sel is made of glass, the wick is fiber- steel heat pipes (horizontal tubes) in ities have been built within the corpo- glass, and the working fluid is a which the working fluidis metallic ration. Table Ilists some of the many fluoridated hydrocarbon. The two sodium. The heat pipes are designed to working fluids and materials used for tubes extending from the pipe are remove heat from the condenser sec- the containment vessels and wick struc- employedtomeasurethethermal tion of a potassium -filled "loop". In tures. Although the name "heat pipe" conductance by means of water- operation, potassium vapor leaving the conveys an image of a right -circular calorimetry. Although the thermal con- last turbine stage of the power system cylinder, they are not limited to this ductanceislarge, the heat pipeis enters the heat -pipe radiator at 1420°F configuration. Heat pipes can be made electrically non-conductive. and exits at approximately the same temperature as a liquid. The heat pipes square, rectangular, oval, spherical, Because of the high thermal efficiency serpentine, re-entrant, or in any con- radiate 50,000 watts of thermal energy, of the heat pipe, substantial savings or the amount required tocondense figuration in which a two-phase, vapor - can be realized with respect to elec- liquid system can exist. The various 216 pounds per hour of potassium from trical power as well as in weight, vol- the vapor state to the liquid state. The heat pipes developed at RCA Lancaster the ume and area. As an example, entire heat -pipe subsystem measures have included many of these shapes radial heat pipe shown in Fig. 4 was and configurations.Fig. 2 shows a only 23 inches by 43 inches and weighs built to cool the collector of a klystron only 17 pounds, about 1/3 the weight "serpentine" configurationthat was for a troposcatter communications sys- used as a heat pipe for cooling an of a comparable conduction -type fin tem. Using a conventional fin -type heat and tube radiator. It was developed for array of integrated circuits. Thispipe 1000 sink, the fan required nearly the Air Force Aero-Propulsion Labora- is made from1/4 -inch copper tubing watts of electricity. When the heat with water as the working fluid. It is tory at Wright Patterson Air Force sink was replaced by the heat pipe Base, Ohio. capable of transferring 100 watts in radiator, adequate cooling was pro- spite of several 180 -degree bends, and vided by a 100 -watt fan. Heat pipes have also been developed demonstrates the inherentflexibility for cooling solid-state devices such as in heat -pipe configuration. Inspacecraftapplications, reduced

45 Table II.List of Symbols SymbolDefinition B Dimensionless constant related to the detailed geometry of the capillary structure R Universal gas constant T. Heat pipe temperature-IC M Molecular weight of working fluid-g L Heat of vaporization-cal/g P (T ) Vapor pressure of the working fluid at the operating temperature , Heat transferred axially along pipe- cal/s e Fraction of wick volume occupied by liquid Length of heat pipe-cm r,, Outside radius of wick-cm g Acceleration of gravity cm/s2 r, Radius of curvature of meniscus at evaporation-cm Fig. 3-Electrically insulated heat pipe. Radius of curvature of meniscus at condenser-cm Radius of vapor space-cm Radius of wick pore-cm Surface tension of working fluidat temperature T.-dynes/cm Wetting angle between working fluid and capillary structure-degrees Angle of the heat pipe to the horizontal-degrees Density of the vapor phase of the working fluid at the temperature of operation g/cm' Density of liquid phase of the working fluid at the temperature of operation- g/cm' Vapor viscosity at operating temperature g/cm-a Liquid viscosity at operating temperature g/cm-s Mass flow rate = Q./L = g/s Fluid wetting angle at evaporation- degrees Fluid wetting angle at condenser-

Fig. 4-Radial heat pipe.

silicon -controlled rectifiers. The heat ating pressure and, because of the must be thick enough to support the pipe shown in Fig. 6, for example, is vapor pressure -temperaturerelation- difference between internal and am- capable of dissipating 1000 watts from ship, will establish a fixed operating bientpressure.Long-term compati- the collectors of two 1N4044 silicon temperature. The temperature of opera- bility tests of the container wall with diodes, while maintaining a 100°C case tion can be pre-set at the time of proc- the working fluid help determine the temperature, with only 90 cfm of cool- essing or modified as desiredifa exact thickness required. ing air supplied by a fan. A comparable reservoir with variable volume is used. conductive -type heat sink would have The wick structure is composed of fine required about ten times the sink vol- Heat pipes designed to provide a con- capillaries. The difference in the diam- ume in order to achieve the same stant temperature have wide applica- eter of the capillary in the evaporator power -dissipation capabilities. tion. As an example, they can be used and condenser sections, coupled with to regulate the operating temperature the fluid surface tension, provides the Itis also possible through a unique of a semiconductor in the presence of driving force for fluid return. How- mode of heat -pipe operation to main- a varying power demand or of a fur- ever, to achieve this force, the working tain constant temperature of a heat nace with varying thermal loads. fluid must "wet" the wick and walls pipe in spite of large variations in of the heat pipe. Fig. 7 shows the effect transferred power. In this operational Design factors of a small -diameter tube placed in a mode the heat pipe varies its access to The heat pipe has been described in fluid bath. An attractive force between the ultimate heat sink so as to maintain right circular form in most of the litera- the liquid and the wall of the capillary the constant temperature. The physical ture. However, as noted previously, tube combines with the surface tension principles involved are the same as there is no theoretical geometric limi- of the liquid to move the liquid surface those by which an air pressure of one tation. In general, because heat is toward the unfilled portion of the tube. atmosphere establishes the boiling added to and removed from the heat The cosine of the wetting angle, 4), point of water as 100°C, i.e.,if an pipe through the container wall by (the angle of contact between the fluid overpressure of inert gas is used within thermal conduction, the wall should and surface) must be less thanir/2.In the heat pipe it will establish an oper- be made as thin as possible. But, it the wick structure the driving force

46 equationforsteadyincompressible flow. For long heat pipes in which rd is much greater than the outside diameter of the wick squared (r.') ,is given by 87110,/ (5) = - r,2)pierp21., The gravitational term, AP, is negative when the evaporationis below the condenser, positive when the evaporation is above the condenser, and zero when the pipe is operated in a horizontal position. The equation for head pressure caused by gravity is given by AP, = piglsina (6) where a is the angle of the heat pipe Fig. 5-Heat pipe panel used in spacecraft applications. with respect to the horizontal position. losses in the system. This relation may For small differences in pressure oc- be expressed mathematically as curring in the vapor -phase region (be- OP, > ap, + .API + (2) tween the evaporator and condenser sections) , the resulting temperature where AP, is the pressure difference of difference, in degrees Kelvin (2K) can the vapor at boths ends of the pipe, be calculated by use of the Clausius- AP, is the pressure loss as a result of Clapeyron equation: viscous drag of the liquid in the capil- Fig. 6-Heat pipe which dissipates 1000 watts from the collectors of two type 1N4044 silicon lary tube, and AP, is the head pressure - (7) diodes. in the liquid caused by the force of MLP(Ti) *I 'T ANC,F ,t CAPILLARY TUBE gravity. There is one other factor which must be considered in the design of a heat The first two pressure losses, AP,. and pipe: The operating pressure of the API, represent two regions of the heat working fluid. Heat pipes can operate pipe:the vapor core and the fluid when the vapor pressure of a work annulus, respectively. The flow of fluid is in the range of 0.01 atmosphere vapor in the central core of the heat Fig. 7-Enlarged view of cap Ilary tube placed to several atmospheres. This pressure in liquid bath to show the effect of capillary pipe, as depicted in Fig. 8, is identical action. range sets the operating temperature to fluid flow through a porous wall by CAPILLARY WICK range for each work fluid. v- CONTAINING VESSEL injection or under suction." This anal- ogy assumes an incompressible liquid, I Z Concluding remarks EVAPORATOR Z '28 laminar flow, and uniform injection or LIOUID The heat pipe has demonstrated its

, suction. Several vapor -flow expressions CONDENSER ( unique abilities to transport and dis- can be obtained based on the magni- sipate thermal energy in a variety of r.t tude of the radial Reynolds number, Fig. 8-Cylindrical heat -pipe structure. sizes, and configurations, and over a R, as follows: is determined by this same wetting wide range of operating temperatures. PrrrVr Currently, heat pipes are being used to angle and the diameter of the capillary. R. (3) - cool semiconductors, electron tubes, In the heat pipe. the capillary action in where V, is the radial -flow velocity at and space power systems. The proven the wick sets up a difference of pres- the channel wall. long life of the sealed heat pipe makes sure between the evaporator and con- it attractive for unattended or inaces- denser ends that causes movement of Based in the two ranges of the radial sible locations. the fluid through the wick. The pres- Reynold's number, the vapor -flow sure gradient can be determined from pressure loss can be expressed as References the expression 471,0,1 I. Grover. G. M., Cotter, P. G., and Erickson. AP, , for R,<<1,and G. F., Structures of Very High Thermal Con- ( cos /1 cos 45 ) rprr, ductance (journal of Applied Physics 3.1, 1964). AP, - 2a (1) (4) r, r, 2. Yuan. S W.. and Finkelstein, A. B., Laminar Flow with Infection and Suction Through a in - 1 - C1,2 Porous Wall (Heat Transfer and Fluid Me- where the terms are as defined ,for Rr>>1. chanics Institute. Los Angeles, 1955). Table II. ap - 3. Knight. B. W., and Mclnteer, B. B.. Laminar Incompressible Flow in Channels with Porous The pressure loss(API)associated Walls (LADS -5309). The fundamental requirement for oper- 4. Cotter,T.P..Theory of Heat Pipes(Los ation of the heat pipe is that the system with fluid flow through the wick from Alamos Scientific Lab Report LA 3246 -MS, the condenser section to the evaporator March 1965). driving force (the capillary pumping 5. Hall, W. B., and Block. F., U. S. Patent Appli- pressure) equal or exceed the pressure sectionis based on the momentum cation No. 418946

47 Ruggedization ofcamera tubes forspace applications J. G. Ziedonis

Television camera systems have played a major part in the exploration ofour closer planets. The heart of these camera systems isthe image -forming. light-sensitive device-the vidicon or the image orthicon. Such tubes must endure the mostsevere environment that electron tubes can be expected to meet: mechanical vibrationand shock during the missile firing and lift-off. As a result,new design concepts have been developed for protection of vidicon tubes. Janis G. Ziedonis Advanced Development Medical Electronics' N SPACE MISSIONSthat use cameras Hoffmann -LaRoche welded, and good welds are crucial for Trenton, N.J. I with vidicons, high -resolution a reliable, strong assembly. received the BS in Mathematics and Physics from video pictures can be maintained only 4) The bulbspacers for electrode sup- Muhlenberg College in Allentown, Pa. in 1959. Fol- if the spacing among internal elec- port must be designed to allow the elec- lowing graduation,he joined the environmental trodes to be inserted in the glass enve- engineering department at the Electron Tube Divi- trodes of vidicon remains the sameas lope without scraping the glass walls. sion. He worked on electron tubes for space appli- before the launch. Vidicons, therefore, Once inserted, these bulb spacers must cations where he contributed to the design and rigidly support the electrodes. If the development of vidicons and image orthicon tubes. are not only treated as light-sensitive He developed tube ruggedization techniques that glass walls are scraped with bulb- electrical devices, but also as mechani- were used on a wide range of conversion tubes In- cal systems. All mechanical systems spacers, particles develop in the tubes cluding the all -ceramic vidicon and the 2- and which eventually get on the target or 3 -inch image orthicons. On a special project for have resonances, and electron tubes faceplate and interfere with tube Jet Propulsion Laboratories, Mr. Ziedonis analyzed are no exception. When an electron operation. the impact phenomena on the Ranger satellite tube is subjected to mechanical vibra- 5) The materials used should have very cameras during their impact on the moon. This was tion, some of its low vapor pressure. The internal elec- accomplished by reviewing and analyzing the last internal electrode trodes operate in a relatively high vac- 800 microseconds of the telemetry data, and per- resonances are excited. At these reso- forming experiments that simulated the last uum, 10-° to 10-' Torr. Any degassing receivedvideosignals.Recently,Mr.Ziedonis nances, tube components may develop occurring after the tubes have been transferred to RCA, Medical Electronics, Advanced large relative displacements which sealed may permanently damage the Development Group, where he is responsible for may deteriorate tube performance or cathode and other electrodes. the development and design of ultrasonic trans- 6) The electrodes must be electrically ducers for medical applications. He is a member produce complete failure. The rugged- insulated from each other. ofSigmaPi Sigma, AmericanAssociationof ization of electron tubes, therefore, in- Physics Teachers, and American Institute of Ultra- 7) The electrode assembly should be sonics in Medicine. volves their redesign. sufficiently rigid to withstand the required vibration and shock environ- 'The RCA Medical Electronics activity has recently Mechanically, the electrode assembly ments. been sold to Hoffmann -LaRoche. represents a complex system with many degrees of freedom. During the Effect of external equipment lems. These resonances are responsible design and ruggedization, two major Resonances of the camera and associ- for most severe mesh microphonics, design concepts must be considered: ated external equipment have consid- loss of resolution, smeared pictures in the internal electrode design and the ef- erable effect on tube performance. Any slow -speed scanning systems, and the fect of external equipment on the tube. kind of resonance in the system that shorter life of the tubes. couples through the tube mounting Internal electrode design Mesh microphonics occur when the severely excites the internal tube -elec- thin mesh vibrates at its fundamental In the design of internal electrodes of trode resonances. In many applications resonant frequency with respect to a a tube, the following properties are at low frequencies, especially below fixed target or photoconductor. The important: 500 Hz, the camera systems have caused video signal is modulated at the mesh 1) The electrically conductive materials poor tube performance. Rattling due resonant frequency. Although the mesh used for electrodes should be antimag- to loose screws, loosely supported mag- resonance in vidicon tubes is well netic because, when magnetic materials netic shields, and lens shutters, ad- become magnetized, they distort the above 2000 Hz, other electrodereso- electron beam path. This requirement versely affects tube performance. nances excite the mesh. In 3 -in, and has only a few exceptions; e.g., the stem Ruggedization techniques 41/2 -in. image -orthicon tubes, the mesh leads need not be antimagnetic. 2) All materials used inside the tube The internal electrode resonances are resonances are below 2000 Hz; these must withstand the 350 to 400°C tube - the major sources of tube performance tubes should not be vibrated at the processing temperatures; annealing deterioration. The thin, long, metal mesh resonant frequencies of 1700 Hz temperature of the materials must be cylinders used in electron tubes have and 900 Hz. well above the maximum tube -processing temperature to preserve the initi- resonances below 2000 Hz. These reso- The most severe environment for elec- ally preselected stiffness. nances have low amplitudes of motion, tron tubes is mechanical vibration. In 3) The metals used must be weldable but they are, nevertheless, responsible most cases, the highest required vibra- because the electrodes are usually spot - for mesh microphonics. The bulb- tion frequency is 2000 Hz. For satis- spacer and the gun -mount resonances factory operation of tubes, bulbspacer Finalmanuscript received October 10,1968. are additional sources of tube prob- resonances should be well above the 48 G 1 -BEAM CONTROL

(6- MESH G5 G4 -FOCUS G3 -DECELERATOR G2 - ACCELERATOR highest vibration frequency; proper bulbspacer stiffness and location must also be determined. Fig. 1 shows a typical, non-ruggedized, STEM one -inch, hybrid vidicon tube. The LEADS TARGET CERAMIC RODS SLABS SLABS analysis of electrode structure of this SUPPORT ELECTRODE BEAD ENVELOPE tube for dynamic environments is per- HEATER E formed with a mechanical system rep- CAT HOCE resenting the tube structure as shown Fig. 1-Non-rJggedized one -inch hiorid vidicon. in Fig. 2. Although the model in Fig. 2 MESH ELECTRODE ASSEMBLY GUN MOUNT is simplified, on the basis of theoretical analysis and experiments performed on actual tubes the following assumptions can be made: 1) Some of the interelectrode supports S LASS (such as the G,-G2 and G.-G.) are ENVELOPE rigid within the 20 -to -2000 -Hz fre- EXTERNAL AP .L1E D:ORCE quency range. 2) Normally, the tubes are rigidly Fig. 2-Simplified equivalent diagram of cr e -inch hybrid vidicon. mounted in the camera. If the tubes BULBSPACEE WELDED are potted in semi -flexible potting com- TO ELECTII)DE pound, they usually resonate below \ N.\ 2000 Hz; therefore, this resonance must TEFLON CMPRE SSE D be included in calculations. BETWEEN E ULBSPACER When non-ruggedized vidicon tubes ANC, GLAS; WALL were subjected to mechanical vibra- --INTERNAL ELECTRODE tion and shock, several major resonances were responsible for complete GLASS TUGE deterioration of the video signals: Fig. 3-Cylindrical bulbspacer for :ne-inch vidicon. 1) Electron -gun -mount resonances:

These resonances were present because _1 NON- RUSGEDIZED of the soft bulbspacers that were initially 4 used to support the electrode assembly. 100 At these resonances, the electron beam was moved out of its electro-optical z paths and loss of resolution occurred. w 0 80 o These resonances were overcome when >4 the heavy glass beads were replaced by -J 1 brazed ceramic discs and the gun -mount D 60 RUGGEDIZED was made considerably lighter. 00 0 1 2) Electrode -support bulbspacer reso- L.L1 nances: Usually, the bulbspacers were 40 overloaded with heavy electrodes which decreased the resonant frequency of the 20 system. When the support bulbspacer resonance was excited, it, in turn, ex- o. cited the mesh resonance. This reso- I 20 IOC 500 11000 2000 nance produced a distorted beam path and video -signal modulation. As a re- FREOUENCv HZ sult, stiffer bulbspacers were designed and, the weight of the electrode assem- Fig 4-Video-signal modulation during sineww.e vibration test of non-ruggedized bly was reduced. These changes as and rJggedized hyb-id vidi ins. 5g peak. shown in Fig. 3 reduced a number of resonances and, in some cases, eliminated them completely. 3) Target -mesh resonances: These resonances modulate the video -signal out- 35 put. As a result, a new mesh processing 30- TUBE technique was developed to reduce the 0 z " 25 mesh microphonics in the tubes. This 0 c process introduced considerably higher 20 RUGGE312ED internal damping in the mesh, and al- TUBE most doubled its resonant frequency to 15

10 4800 Hz. This design change enabled _J the mesh to withstand the most severe o 5 environmental requirements suchas I

3000 -g -peak, 0.5 -ms sine -pulse shock. .5 I 1.5 2.0 25 0 3.5 4.0 4.5 5.0 Vibration frequencies at which the FRONT END OF -HE II BACK END OF THE T.J5E non-ruggedized tube performance was TUBE (TARGET -MESH) (HEATER -CATHODE) Fig. 5-Weight distribution curve for non-ruggeiized and ruggedized one- nch vidicons.

49 considerably distorted are shown in Fig. 4.

A The electrode weight distribution Ai Ai Ai Ai - - - 141 A 1.1 curve is shown in Fig. 5. This curve is 7 0 411 e - used to determine the required stiffness 0 .4 14 " of the bulbspacers and to assure that el 0 . 4,4 01 - the electrode supports resonate well above the required vibration frequency ( range. This diagram was obtained by 0.24" carefully weighing each electrode, by 1.23" measuring its length, and then compar- Y- ing the weight -to -length ratio to the 2.66 - tube length. For the non-ruggedized 2.741 tube, the data indicate a non -uniform SUPPORT ON WW1 3.95 --""' BACK END OF electrode mass distribution. .90 TUBE (HEATER -CATHODE) With the assumption that the damping FRONT END OF TUBE SUPPORT ( TARGET- MESH) within the metal bulbspacers is negligible, the required bulbspacer stiffness Fig. 6-Equivalent diagram of the electrode assembly. Kr may be obtained as follows:

BIMETALLIC BULBSPACER KT = CAW (I) If the desired electrode assembly reso-

ELECTRODE nance is 3000 Hz, the required bulb - ASSEMBLY spacer stiffness is 73.1x104 lb./in. This WHEN ASSEMBLY IS HEATED THE BULBSPACER DEFLECTS magnitude suggests a very stiff bulb - TOWARDS THE ELECTRODE spacer assembly. The total stiffness, ASSEMBLY. of each of the selected bulbspacers is

Kr = + + + ... (2) According to Fig. 5, another set of Fig. 7-Sectional drawing of bulbspacer. bulbspacers is needed. The location of the new set of bulbspacers is calculated with the use of Fig. 6 as a reference and the following static equilibrium equations: = 0 (3) BRAZED CERAMIC SPACERS = 0 (4) Fig. 8-RCA ruggedized hybrid one -inch vidicon, 8567. From these equations, the location of the third set of bulbspacers was cal- ,-ELECTRODE CONTACT LEADS culated to be 1.92 inches from the back end of the assembly. Gs 53 G4 G3 G2 111.H -It

Symbols

F static force lb KT total electrode lb/in support spring TARGET stiffness M mass (W / g) lb-see/in Fig. 9-All-ceramic one -inch vidicon. Mo moment for staticlb/in conditions

co angular frequencyrad/sec LAUNCHEDMISSILES (27r/) 10 AIRCRAFT vibration frequency Hz Several new design concepts were HELICOPTERS developed to overcome the existing S / manufacturing and assembling difficul- HELICOPTERS WITH EQUIPMENT 014, ties concerning bulbspacers. For ex- 2 VIBRATION ISOLATORS T --- ample, the stiffness of a bulbspacer is

0 I generally controlled by itsmaterial 20 SO 100 200 500 10002000 configuration, thickness, and length. FREQUENCY Hz With stiff bulbspacers, the electrode- Fig. 10-Sinusoidal vibration test levels per MIL -STD -810. assembly insertion problem in the glass

SO LAUNCHED 11331LES 10g RMS (APPROX cylinder could be simplified with the evaporated on the inside walls of the 03-- use of bimetallic materials for the bulb - ceramic tube envelope. The ceramic spacers. The bimetallic material is used cylinder itself resonates at 5000 Hz. in such a way that when the gun -mount This design technique eliminates the is preheated prior to its insertion in the need for metal cylinders and bulb - 10 20 50 100 500 1000 2000 glass envelope, the bulbspacers deflect spacers. Fig. 9 shows the all -ceramic FREQUENCY HZ towards the electrodes and insertion one -inch hybrid vidicon built for the Fig. 11-Rs ndom vibration test levels per MIL -STD -810. 60 becomes easy, as shown in Fig. 7. As Jet Propulsion Labs, Pasadena, Cal.' OPERATIONAL NONZPERATIONAL a result, the glass walls are not scraped AL.. CERAMIC HYBRID I" VIDICCN -BEING DESIGNED and no particles are generated. During Requirements and capabilities 50 the stem -sealing process, the glass With the growth of military and space 40 softens near the gun -mount and the exploration programs, the environmer.- plain, stiff bulbspacers push the glass tal requirements for the components 30- out and deform the glass walls; how- have been increasing. Figs. 10 and 11 S&L MAGNETIC I" VIDICON 72i'3A ever, stiff bimetallic bulbspacers do not show vibration specifications for com- 20- cause deformation. ponents such as vidicons and image orthicons, as required by the military NYBRID VIDICON TYPE4462%ID OTHERTYPES Another type of bulbspacer that was 10 - M AGNET IC - 7 2 63 A t - -..._.... developed and is widely used in vidi- contracts. These environmental re- 1

_I 1 1 quirements may not seem too severe. 1 11111111 I 11 1111 con tubes for military and space proj- 20 50 100 500 1000 2000 ects is the cylindrical bulbspacer that but, unfortunately, several major resonances exist in the systems using these FREQUENCY HZ compresses thin sheets of tetrafluoro- Fig. 12-RC-A vidicon tube capabilities-sinewave vibration. ethylene (Teflon -a trade name of electron tubes. These resonances subject the tubes to vibrations levels which OPERATIONAL Du Pont de Nemours, Inc.) between ...: NONO AAAAA IONAL the glass and the bulbspacer, as shown exceed the design specifications and, : CERAMIC WIRRIO I" 1510011 SO 1012__IL_,10rj thus, cause failures. With complete L0 BEMS 0EvELOPED in Fig. 3. Teflon prevents the metal i NYBR10 I" VIDICON TYPE 4452 evaluation of the system, these failures S 4:V from scraping the glass during the ALL 666666 IC 1/2" VIOIC3N 4427 2501MS\ 4 electrode assembly insertioninthe can be prevented. Figs. 12 and 13 sum- ./ ALL MA6RETIC I" VIDICOM 'MA 2001115 \ marize thecapabilitiesof one -inch 117111910 1.-VIDICON TYPE AMU 175111117 glass cylinder. The complete electrode 0.1- and bulbspacer assembly withstands vidicon tubes that have been ruggedized in the last four years. The image- HYBRIDI" VIDICON TYPE 1287 0911111, the tube -processing temperatures. ; orithicon tubes, being more compli- 20 50 100 500 1000 2000

A combination of all the above design cated, have not reached as high FREQUENCY Hz features considerably improved the environmental capabilities as some of Fig. "3-RCA vidicon capabil tins during random performance of the hybrid vidicon the vidicon tubes. Figs. 14 and 15 show vibration tes:s. 'ON -OPERATIONAL tube (Fig. 4) . the presently available image -orthicon (PERATIONAL This type of design improvement was tubes that have reasonable environ- 20 2 1 0 incorporated inother vidicon tube mental capabilities. types, such as the all -magnetic, all - Conclusions electrostatic, reverse -hybrid. As an 3 I 0( 7 991 example, Fig. 8 shows a ruggedized During the ruggedization of camera to one -inch hybrid vidicon. tubes, not all the electrode resonances have been eliminated. However, the Image -orthicon tubes are facing similar major and most damaging electrode environmental requirements, but the structures have been eliminated, as solutions to the design improvements shown in Fig. 8. With the ruggediza- are more difficult. All components are tion of camera tubes, the number of proportionally larger and heavier, and particles in the tubes was considerably 2 the tube processing is more severe. The reduced, the weak electrode structure C 2000 choice of available materials that could was removed, the operational tube per- 20 50 100 200 500 1000 FREQUENCY I-z be used in these tubes is reduced be- formance during severe environments F g. 14-RCA image orthicor capabilities during cause of the higher vacuum and the was considerably improved, and higher sinewave vibrrion. sodium, potassium, cesium, and anti- reliability was achieved. NON -OPERATIONAL OPSRATIONAL _ mony elements used for making photo- IS S RMS cathodes. These tubes are also more References I. Ochs. S. A.. Development of a Sterilizable Rug- sensitive to particles which, from pre- gedizedVidiconforLunar and Planetary t o vious observations, have torn both Photography. Final Report for jet Propulsion Laboratories, Pasadena. California, Contract 2 i 0 10 g RMS glass and magnesium oxide targets. No. 950985.

2. Ochs. S. A.. and Marschka, F. D., "A Steriliz- CC With the glass -envelope type of tubes, able and Ruggcdized Vidicon," this issue. 031-- it is impossible to have all the internal 1. Harris and Crede. Shock and Vibration Hand- book. Vols. I and II (McGraw-Hill Book Co.) tube components rigidly mounted. 4. Hartog. Mechanical Vibration (McGraw-Hill Therefore, a new design technique was Book Co.) 61 5. Ziedonis.I.G.. "Ruggedization Techniques 3 I I X 01 1 111111 _L developed in which electrodes are for Electron Tubes," private correspondence. 10 20 50 100 500 1000i1 2000

FREQUENCY HZ Fig. 15-RCA image orthicon capabilities during random vibration test. The coaxitron J. A. Eshleman B. B. Adams

The Coaxitron integrates the complete RF input and output circuits and the electron - interaction circuits within a common vacuum envelope. This design eliminates the need of ceramic -window exposure to extremely high RF voltages and ceramic -to - metal seal to extremely large RF currents. It eliminates external spring-conta:t fingers, blockers. and tuning. By integration of RF circuit elements into the tube envelope. the high-RF-voltage portions of the resonant circuits are insulated by a vacuum dielectric which has a high dielectric strength and excellent self -healing characteristics. Stored energy in the resonant circuits and in the lead inductances is reduced, and the dielectric dissipation and the propensity to arc over are minimized. Because 'f these advantages. the Coaxitron RF circuit components can be designed for maximum amplifier performance. Instantaneous bandwidths in excess of 10% at the megawatt RF power output levels are readily feasible. These integral packaged amplifier tubes resulted from an extensive developmental effort in the Super Power Devices Group at Lancaster. This paper describes the design features of the RCA Dev. No. A15193B Coaxitron. THE COAXITRON CONCEPT, as developed in the RCA Dev. No. A15193 (Fig. 1), meets the require- 4g41 ments of chirp signals and state -of -the - COAXIAL OUTPUT TRANSMISSION LINE art, multi -channel, megawatt -level, /AND CERAMIC WINDOW radar systems. The design require- are listed in Table I and a simpli- //mentsfled longitudinal cross-section is shown ELECTRONIC G ::TTER /ION PUMP AND in Fig. 2. For ease of discussion, it can MAGNET be considered as consisting of three basic areas: the triode electron -interaction structure, the output circuit, and the input circuit.

ELECTRONIC STRUCTURE, Triode electron -interaction BLOCKER, AND ANODE Fig. 3 shows the electron -interaction structure of the Coaxitron. It is comprised of a cylindrical array of 48, essentially independent, grounded -grid unit triodes. Each unit triode has a directly -heated, matrix -oxide filament, a double -wound grid, and a high -dissipation anode. A total active cathode area of 150 square centimeters was provided so that the required emission could be obtained with only modest grid -filament RF voltages. The active filament emitting surface is ANODE SUPPORT spaced approximately 0.015 in. from CHOKE the inner layer of grid wires. For maintenance of this spacing under hot and cold filament conditions, each filament is suspended at one end with a unique pantograph system. These systems take up the expansion of the filament when itis heated to its operating temperature and thus maintain the grid -filament spacing. INPUT DRIVE LINE BLOCKER (J-15073) The filament cross section is reduced near the support ends. This section is

Fig. 1-RCA Dev. No. A15193B Coaxitron. Final manuscript received October 21. 1968.

52 cUTPUT CERAMIC A INixEw

BROADBAND OUTPUT CIRCUIT

RIBA [IBA ND INPUT CIRCUIT

Fig. 2-Simplified longitudinal cross-section of the Coaxitron.

CAIuntil

ps+441, #.0411 -1-

111,

(Si TPuT REGION - -Ligon COOLANT COOLANT COURSE

Authors B. B. Adams (left) and J. A. Eshleman stand alongside a Ccaxitron being readied for shipment. GRID `GRID ANODE WIRES BLOCK J. A. Eshleman and step transitiors for the ORSAY Klystron Appli- Fig. 3-Triode electronic structure. Product Development cation andisdoing cavity design work for the Industrial Tube Division RCA Dev. No. A 1401 Klystron. Mr. Eshleman is a COAX IA I. OUTPUT Electronic Components member of the Sigma Pi Sigma. TRANSMISSION LINE Lancaster, Pa. received the AB in Physics from Franklin & Mar- shall College in1961 andispresently pursuing B. B. Adams .1 I I, the MS in Physics at the same school. Mr. Eshle- Super Power Tube Design Engineering I II.C.111 Industrial Tube Division man joined the Super Power Tube Design and OUTPUT COUPLING Development group atRCA, Lancasterin1957. Electronic Components CAPACITOR From 1958 to1961he was responsible for the Lancaster, Pa. testing of developmental super power tubes and received the BSME from Pennsylvania State Uni- for test equipment maintenance on BMEWS, BTL. versity and is currently pursuing a course of grad- and TRADER. From July 1961 to 1963, as a Product uate studies leading to the MS in Engineering. Mr. Adams joined the RCA Electron Tube Division at AUXILIARY OUTPUT Development Ergineer, he established a computer CAVITY programforthe outputcircuitand RF bypass Lancaster in 1957 as a design and development capacitor design for RCA Dev. No. A -2696A. He engineer. He has been instrumental in the mechanical design of RF circuitry, cold probe test gear also was responsible for the co,d probe ad,ust- RE-ENTRANT INDUCTIVE COUPLING ments on the output circuits and for the evalua- and special devices utilized in the development of DC VOLTAGE: SPOKES a variety of super power tubes and circuits. He has BLOCKING tion tests. From 1963 to 1966, he was responsible CAPACITOR for computer design calculations on the output also provided consultation and engineering sup- circuits of RCA developmental coaxitrons. He also port for various Government research and develop- made cold probe measurements and computer cal- ment projects. As Project Engineer, Mr. Adams was OUTPUT LOAD END CAVITY culations on coaxitron input circuits. From 1966 to responsible fo the mechanical design of the RCA 1967, he was the Project Engireer for the RCA Dev. No. A-2696 and A-15193 coaxitrons. He has Dev. No. A -15'91B coaxitron and was responsible also been responsible for the design of a variety for the design calculations and configuration. Mr. of high power cavities. He is a licensed profes- Eshleman is presently working on a compute' pro- sional enginee in the Commonwealth of Pennsyl- 11...... ------"--...... ELECTRON Pi.Pi T2 u Sigma INTERACTION gram to calculate electron trajectories in beam type vania, a member of Tau Beta REGION tubes. He isalso designing directional couplers and the ASME.

Fig. 4-Broadband out- IN put circuit. 53 INPUT SLAVE END CAVITY called the "filament lead correction"; an RF power output in excess of 1 its purpose is to minimize heat loss to megawatt (MW) with a ripple of less GRID -CATHODE ACTIVE INPUT the filament supports and to provide than 1 dB over a bandwidth of 34 MHz REGION a constant temperature in the active centered on approximately 435 MHz. emitting region. The filament tempera- Because of the many variants associ- SAPPHIRE BALL ture in the Coaxitron is maintained ated with the design, the first prototype within ± 7.5°C over the active 4 -in. design was developed with the use of SHUNT CORRECTING STUB length. computer techniques. In this manner, FIRST QUARTER the effect of slight changes in any of WAVE MATCHING A double -wound grid of 0.0033 -in. - A NODE LINE SECTION the variables was studied very eco- diameter tungsten wire is used in each nomically. As a result of this approach, CHOKE BUCKET triode unit for electrostatic isolation of LINE SECTION a design closely approximating the BROADBAND the filament and anode and for class -B INPUT final tube was derived in a relatively TRANSFORMER tube operation with zero grid bias. This design eliminates the need for short time. SECOND QUARTER WAVE MATCHING neutralization and the need for DC iso- Fig. 4 shows the basic output circuit LINE SECTION lation between the grid and cathode. used in the Coaxitron. This design has two resonant circuits which are over - The tungsten grid wires for each triode coupled through the "inductive cou- unit are supported on OFHC copper INNER CERAMIC pling spokes." The first resonant circuit OUTER CERAMIC VACUUM SEAL lands which extend radially outward VACUUM SEAL comprises the "output slave -end cav- from a central cylinder called the "grid 1 0.11M DRIVE LINE ity," the "electron -interaction region," cylinder." The grid wires are fastened the "output load -end cavity," and the to these lands by a special notching Fig. 5-Broadband input circuit. "re-entrant DC voltage -blocking capaci- and peening technique which permits tor." The second cavity comprises the the outer grid wire to be hidden from "auxiliary output cavity" and the area 1300 . direct cathode emission by the inner below the large diameter step just grid wire. Because the grid spans are below the "output ceramic window." short and the copper lands form an Output coupling to the over -coupled 1200 excellent heat sink for heat generated circuits is made through the "output by grid interception, the design exhi- coupling capacitor" into a 50 -ohm, dB bits excellent mechanical stability. iv 1100 31/2 -in. -diameter coaxial transmission The output of the unit triodes is par- line. The output ceramic window, like the input seal, is located outside the 1000 . tially formed by a doughnut -shaped anode. The anode is supported only on resonant circuitry; therefore, it is sub- 410 420 130 440 one end by three metallic support jected only to the relatively low RF 131Qt FM')- 0111, columns spaced 120 degrees apart. voltage on the output transmission Fig. 6-Power ou put versus frequency. Two of the support columns conduct line. This arrangement minimizes arc- anode liquid coolant. The flow of cooling and dielectric losses in the window ant through these courses feeds the area. fluted heat exchanger on the back of The "re-entrant DC voltage blocking the bombarded anode surface. The DC capacitor" is used only as a means of voltage for the anode can be applied by DC isolation of the anode from the tube any one of the three support columns. envelope. With this isolation, a DC an- Because the support columns are me- ode voltage of 20 kV can be applied tallic, they form coaxial lines which without arc -over. Ideally, the RF im- may leak RF energy through the tube pedance between the tube envelope envelope. Such leakage is undersirable and anode is negligible at the output because it represents wasted energy, slave -end and load -end cavity. There- presents a potentially dangerous RF en- fore, the circuit is readily recognized when an RF short is envisioned be- I - NO vironment to operating personnel and U increases the potential of RF interfer- tween the anode and tube envelope at fs ence to nearby equipment. Therefore, these points. the RF leakage from each support col- Input circuit C umn is attenuated to acceptable and safe levels by special RF chokes exter- 40 -- - The input circuit of the RCA Dev. No. nal to the tube envelope (Fig. 1). A15193B transforms the impedance of the 50 -ohm, 15/8 -in. coaxial input transmission line down to the impedance 410 420 430 440 Output circuit of 0.5 ohm at the driven end of the 1-111 tj1 - 1111, The output circuit for the RCA Dev. Fig. 7-Efficiency and power gain versus electron -interaction structure. Despite frequency. No. A15193B was designed to deliver impedance transformation require-

54 ments of approximately 100:1 and var- seal. After a series of experimental mately 0.58 dB. Fig. 6 also displays ious complicating factors such as the tests, vacuum -potted Dow Corning the typical band-pass characteristics of DC isolation of one end of the filament Sylgard was selected as a substitute the A15193B. because it featured a high dielectric structure, an input circuit design was Fig. 7 shows the efficiency and power strength, eliminated trapped air, was derived for the initial RCA Dev. No. gain as a function of frequency. The semi -flexible, and provided excellent A15193B tubes which gave an input fall -off of power gain near the upper VSWRof less than 2.5:1 over the 34 - metal bonding characteristics. The edge of the operating band is caused MHz band. Computer techniques were chokes so fabricated exhibited excel- by a greater increase in theVSWRof the again utilized to develop the input cir- lent reliability. input circuit at the upper band edge cuit design for the first prototype. The massive doughnut -shaped anode than at the lower band edge. Fig. 5 is a simplified drawing of the with its associated support structure Coaxitron input circuit. Two quarter - was required to withstand 18, 15-g. Table I-Design requirements for RCA wave coaxial sections are utilized for 11 -ms shock pulses without arc -over, Dev. No. A1519313. transformation of the 50 -ohm input loss of vacuum integrity, or permanent Electrical impedance to the approximate 0.5 -ohm distortion. A support arrangement of Frequency 418 to 452 MHz Pulse width 13 as impedance at the electron -interaction three stainless steel pipes, connected Bandwidth 34 MHz region. Because of the extremely close as parallel cantilever beams, was Uniformity of response 1 dB (max) Power output 1.000 kW (min) spacings required in the first quarter - selected for the anode after analyses Efficiency 40% (min) wave matching line section, small sap- indicated that the structure would Power gain 13 dB Input VSWR 2.5: 1(max) phire balls, used as dielectric spacers, withstand the specified environment. Plate voltage 20 kV Mechanical maintain the coaxial spacing and The results of these analyses indicated Weight 175 lb.(max) alignment. that a maximum anode displacement Length 39 in. (max) Diameter 17 in. (max) Since the A15193B was designed for of 0.016 -in. would occur during the Environmental -54 to +55°C zero -bias class -B operation, it was feas- shock pulse and that this displacement Temperature would not cause arc -over nor a perma- Altitude 15.000 ft. ible to connect one end of the filament Shock 15 g at 11 ms (half nent offset in the anode structure. For sine wave) Dc -wise to the grid structure via the Vibration "input slave -end cavity." The "input assurance of vacuum integrity, the sup- Frequency 5 to 500 Hz port pipes were connected to their Amplitude 0.100 in. from 5 slave -end cavity" was used in this ar- to 65 Hz respective high -voltage Pc ceramic in- 2 g from 65 to 500 Hz rangement to provide grid -cathodeRF Acceleration sulators through shortOFHCcopper isolation and to insure that anRFvolt- age antinode would occur approxi- strain -isolation rings which provided Table II-Typical operating conditions of mately in the center of the active input a resilient, but strong connection. RCA Dev. No. 15193B. Pulse plate voltage 16 kV electron interaction region. The weight of an 18 -pound ion pump Pulse plate current 157 A magnet connected externally to the Pulse drive power 45 kW Because one end of the filament was tube presented severe mounting and Pulse Power Output 1,070 kW connected Dc -wise to the grid structure. Efficiency 42.5% support problems. Any large relative Power gain 13.8 dB DCandRFiso- 800 A (steady state) it was necessary to both displacement of the magnet with re- Filament current late the opposite end. This function is Filament voltage 1.38 V (steady state) accomplished with the quarter -wave spect to the associated ion pump dur- "choke bucket line section" and the ing the shock pulse could destroy the Summary "inner ceramic vacuum seal." The ion pump, the pump tubing, and the Coaxitron vacuum integrity. Various The RCA Dev. No. A15193B Coaxi- "choke bucket line section" not only in- magnet support designs were investi- tron was successfully designed. fabri- RFleakage at the "outer ceramic hibits gated and many proved bulky and im- cated, and evaluated for wide -band, vacuum seal" but through a 1:1 trans- high -power operation. This Coaxitron formation via half -wave coaxial sec- practical. A simple solution was ultimately obtained by use of a magnet is designed for grounded -grid zero -bias tion, termed the "shunt correcting bolted directly to the Coaxitron hous- class -B operation and provides a typi- stub," it provides a high grid -cathode ing through a Delrin plastic pad. This cal power output of 1 MW, a power impedance at the base of the "grid - system reduced the relative motion of gain of 13.8 dB, and a conversion effi- cathode active input region." The "in- the pump and magnet to negligible ciency of 42.5%, with a 1 -dB instan- ner ceramic vacuum seal" serves a dual values and thus protected the tube taneous bandwidth of 34 MHz centered role as a coaxial input window and from being damaged by the magnet. on approximately 435 MHz. as aDCinsulator for the "hot" filament terminal. The elimination of moving circuit com- Test results ponents and the placement of ceramic Mechanical design The performance of RCA Dev. No. windows in low -voltage fields combine A15193B met or exceeded all the de- with the use of welded and brazed Although an oil dielectric had been sign requirements listed in Table I. joints to reduce the tube -circuit weight, used in anode chokes of prior Coaxi- Typical operating conditions are listed improve tube capability for withstand- tron tubes, oil could not be used on in Table II. Fig. 6 shows the variation ing rotatively severe environmental the A15193B because of its tempera- of power output across the frequency conditions, substantially reduce main- ture expansion characteristics and its band with constant plate voltage; the tenance problems, and increase tube tendency to develop leaks at the choke uniformity of response is approxi- reliability.

55 Noble -gas -ion lasers R. J. BuzzardJ. A. PowellJ. T. MarkH. E. Medsger

RCA has been manufacturing noble -gas -ion lasers as a commercial product since early in 1967 when a 100-mW argon -ion laser was first designed by K. G. Hernqvist: Since then, the line of commercially available ion lasers has expanded to seven types ranging in rated output power from 10mW to 10W and utilizing such inert gasesas argon, krypton, and neon. The product line includes both pulsed and continuous -wave types and covers the visible, ultra -violet and infrared portions of the spectrum. This paper describes the major design features and performance of noble -gas -ion lasers.

J. A. Powell equipment, and product design. He was responsi- for use in ultra -high vacuum applications and laser Heat Transfer Engineering ble for the design of several large space chambers. applications. He has been granted eight patents Industrial Tube Division He was also responsible for the mechanical design and has two additional patents pending in elec- Electronic Components, Lancaster, P8. of a 5 -MW peak, 300 -kW average power cavity for tron tube design, lasers, and vacuum technology. receivedthe BSinCeramic Engineeringfrom the RCA 2054 family of Super Power Tube and com- He has presented and published numerous papers. N.Y.S. College of Ceramics, Alfred University. He plex air and high purity water systems. As engi- He is a member of the AVS committee on stand- joined RCA Lancaster in 1955 as a Product Devel- neering leader of the Super Power Tube Vacuum ards,IEEE,IVOST, ASLE,IES, and the ASTM opment Engineer in the Small Power Tube Design Group, he was responsiblefor the mechanical Committee E-21for standardsinthe aerospace section where he worked on the development of design and development and the initial production industry. small ceramic vacuum tubes. In 1957, he was as- of the LD2100, 100-mW argon gas laser and the signed to a specialproject team developing bake - LD2101 1-W argon gas laser. Prior to his present able ceramics, sapphire, and glass -to -metal seals position, he was Manager of Production Engineer- R. J. Buzzard, Ldr. for use in a large, ultra -high vacuum chamber de- ing, Super Power Tube ITD. Mr. Medsger has com- Heat Transfer Devices Engineering signed specifically for the military. From 1961to pleted mechanical engineering courses at Syracuse Industrial Tube Division 1965, Mr. Powell was assigned to the Vacuum Engi- University and Penn State University and industrial Electronic Components. Lancaster, Pa. neering Group and engaged in the design and de- design at Maryland Institute and Johns Hopkins received the AB in physics from Franklin and Mar- velopment of ceramic -to -metal and glass -to -metal University. He is a member of ASTME and is past shall College in1962. He joined RCA in 1953 in feedthroughs including a complete product line of chairman of the Lancaster Chapter .t'89 ASTME. the Large Power and Gas Tube Laboratory. Upon bakeable sapphire and fused -silica windows for completing his undergraduate work, Mr. Buzzard ultra -high vacuum applications. Since 1967,Mr. was reassigned as an engineer in the Thermionic Powell has been responsible for laser -tube design J. T. Mark, Mgr. Converter Engineering Group where he participated Regular Power Tube Engineering at Lancaster. Before joining RCA, Mr. Powell was inthe successful development ofRCA'sinitial with the General Electric Company and with the Industrial Tube Division nuclear -fueled converter. In a continuation of this Star Porcelain Company for two the Electronic Components, Lancaster, Pa. effort, he developed the basic computer programs Frenchtown Porcelain Company for five years as a received the BS in Radio Engineering from Val- required to solve the thermal and electrical design Research and Development Engineer working in the paraiso Technical Institute in 1950 and joined RCA problems for high power converters and their asso- field of ceramics engineering. Mr. Powell has been Lancaster the same year. Since 1950, he has done ciated components. Since 1963, he has extended granted one patent in the field of ceramics and has graduate work inphysics and has had extensive this computer approach to a comprehensive equa- authored several papers on the subject. He isa experience in the design, fabrication. and testing tion which analyzes the design for all the various member of the American Ceramic Society and the ofultra -highvacuum equipment andsystems, Parameters. Mr. Buzzard was engaged in RCA's American Society of Metals. megawatt electro-mechanicaldevices,megawatt original investigations in power conditioning asso- vacuum tubes and gas lasers. Mr. Mark was as- ciated with thermionic converters. He participated signed initially to the design and development of in the development of tunnel diode inverters and H. E. Medsger, Supt. black and white and color kinescope tubes. Upon their evaluation in circuits employingthermionic Manufacturing Dept. 953 transferto theSuper -Power TubeEngineering converters.Mr.Buzzard was promoted to Engi- Industrial Tube Division Department, he became responsible for the design neering Leader in 1965. During 1965 and 1966. he Electronic Components, Lancaster, Pa. and fabrication of one of the first multimegawatt providedthetechnicaldirectionforthe NASA joined RCA in1958 as an associate design and test equipments at the Lancaster plant. He also IsotopeThermionicDevelopmentProgram.Cur- development engineer withtwenty-two yearsof designed and developed new electron gunsfor rently,Mr. Buzzardisdirecting the development diversified product design and manufacturing engi- super -power microwave tubes. Assigned as Project of a number of special power devices; these in- neering experience. He had been associated with Engineer on the Matterhorn Projectin1958, Mr. clude heat pipes, alkali metal vapor arc lamps, and the Glenn L.Martin Co., as manufacturing engi- Mark was responsible for the overall direction and high -power gas lasers.Mr.Buzzard has co-au- neer. IBM as assistant project engineer on com- control of the design and development of the ultra- thoredseveraltechnicalpapers onThermionic puter mechanisms: General Electric TV Division as high (10-10 Torr) vacuum system for the C-Stel- Conversion Devices. a manufacturing engineeringleader; and P.G. laratorresearchfacility.Later,he directedthe Spaulding and Associates as executive engineer. design of several large vacuum systems. Promoted Since joining RCA he has been engaged inthe toEngineeringLeaderin1961.Mr.Mark was rr HE BASIC COMPONENTS of a DC - designdevelopmentand fabricationofSuper responsible for the conception. design, and devel- Power Tube evaluation equipment, manufacturing opment of sophisticated components and systems 1 excited gas laser are the discharge tube, a stable optical resonator, and the DC power supply. Fig. I shows a cross- sectional diagram of a typical gas laser. The discharge tube contains a cathode, an anode, and a long plasma -confining bore structure. The plasma bore is the region in which the coherent light is generated. The bore can be formed from either a uniform insulator tubing, or a series of short tubing sections uniformly spaced and insulated from one another. The discharge tube is filled with the noble gas at the proper pres-

In the photo (left to right) Powell, Mark, Medsger, and Buzzard are seated behind some of 56 the Lancaster -produced lasers. sure (typical in the order of 0.1 torr). Gaussian distribution of intensity Energy is supplied to the plasma by a across the beam diameter. A Gaussian FULL REFLECTOR discharge current between the cathode distributionrequires that thelaser MIRROR BORE CATHODE OUTPUT BEAM and anode. The optical resonator con- operate in the fundamental transverse ANODE sists of two mirrors, one totally reflec- mode (TEK.,). Mode selection is active and the other partially reflective, complished by the proper choice of placed at opposite ends of the dis- mirror radii and the bore diameter or BREWSTER'S IONIZED GAS PARTIAL REFLECTOR charge tube. When this optical cavity is aperture in the optical resonator. Figs. ANGLE WINDOWS MIRROR properly adjusted with respect to the 5 and 6 show the diffraction losses for Fig. 1-Typical gas laser. discharge tube, "lasing" takes place as the two lowest -order (TENOR, Temo,) the emitted light energy is coupled out modes of a stable resonator using a pair GRAPHITE BORE through the partially reflective mirror. of identical mirrors.' The correct com- SEGMENTS A magnetic field is usually provided bination of mirrors and bore diameter ANODE around the discharge tube to improve at these low -order modes will result in the power output. substantial losses in light output when the laser is operated at higher -order CAT RODE Design considerations modes; i.e., the higher -order modes are MOUNTING RING The five major components that affect attenuated and do not appear in the the performance of noble gas ion lasers laser output. Fig. 2-The RCA-LD2110 10-mW Argon ,aser tube. are:the bore structure, the optical cavity, the envelope, the electrode Discharge -tube envelop* structure, and the power supply. The discharge -tube envelope is made of fused silica(quartz), which has Graphite bore structure many desirable properties for this application. The material is capable of RCA continuous -wave ion lasers utilize containing the graphite bore segments a segmented graphic bore structure', of at temperatures as high as 1000°C. In the type shown in Fig. 2. This bore addition, the infrared -transmission structure was chosen because it exhi- properties of fused silica allow the heat bits a minimum of sputtering and bore from the graphite segments to be ra- erosion. Fig. 3 is a photograph of an diated to an external heat acceptor individual bore disc which is composed with a minimum of absorption. The of high -purity, isostatically pressed continuous -wave lasers generally have graphite. The material was chosen quartz windows fused permanently to after extensive testing and comparison the tube structure to provide a com- Fig. 3 --Graphite bore segment. with other possible bore materials. The pletely vacuum -tight, bakeable struc- outer diameter of the graphite discs is ture. The windows are placed at determined by the input power level of Brewster's angle with respect to the the laser. The center bore hole in each incident light beam to minimize losses disc is between 1 and 4 mm in diam- caused by reflections at the window eter. The exact bore diameter is de- surfaces. termined by opticalconsiderations. The length of the bore determines the total gain of the laser and the total Electrodes light output. The cathode material is also very im- MI u, portant becauseit must have good Fig. 4-Spherical mirror resonator. Optical cavity electrically emissive properties, withstand high ion bombardment, and be Ballast tank The optical cavity consists of two resistant to gas impurities that may be During the operation of the discharge dielectric -coated mirrors mounted in a evolved from other tube components stabilized tuning mechanism. The ori- during discharge operation. In the tube, small amounts of gas are grad- entation, reflectivity, and geometry of ually lost as a result of sputter pump- high -power -laserdischargetubes,a ing. The phenomenon is referred to the mirror all affect the performance barium -impregnated tungsten -matrix of the laser. The mirrors are, in most as "gas clean-up" and results when cathode is used. Emission life in excess cases, curved and arranged as shown atoms sputtered from the bore confine- of 10,000 hrs can generally be expected ment structure "bury" gas atoms on in Fig. 4. The condition for resonant from this type of cathode when used stability of a laser is 0 < g, g < 1 for cooler surfaces. Although the sputter- g, = (1 -d / R,) and g, = (1 - d/R,) in the laser environment. ing and the resulting gas clean-up are where d the distance between the mir- Because the anode material has little minimal is a well -processed graphite - rors, and R, and R2 are the radii of the effect on laser performance, it is gener- bore discharge tube, provisions must respective mirrors. ally made of the same material as the be made for extra gas storage. Conse- Many laser applications require the bore segments and made simply an ex- quently, a gas ballast tank is provided output beam of the laser to have a tension of the graphite structure. in one of the configurations shown in

57 100 90 TEM.. MODE 60

20 CONCENTRIC BULB TYPE 10

d 8 a 6

g 4

1...g. 2 TV BULB 1.0 TYPE

0.6 as

0.2

0.1 01 0.2 0.3 04 10 2 4 6 910 20 3040 00 N 02/ Xd Fig. 5-Power loss per transit of the fundamental (TEM00) mode for circular mirrors.

100 80 TEM. MODE 60

CONCENTRIC TANK TYPE

Fig. 7-Ballast tank configuration. 0.4

0.2

0.1 01 02 0.4 0.6 10 2 4 6 10 20 4060 100 N 2/x Fig. 6-Power loss per transit of the TEM,e mode for circular mirrors. ,00 132 r--

131,

#010 0

126

0 010 125 1.0 oo 0 2.5 3.0 3.5 4 0 4.5 TOTAL POWER INPUT (Awl OC EXCITATION (AMPERES) Fig. 9-Light output versus power input for Fig. 10-Mirror mount. Fig. 8-Typical excitationcharacteristics. Argon -ion lasers.

58 can be achieved through the use of an axial magnetic field, which helps to confine the plasma within the bore structure and also reduces sputtering and bore erosion. The field is supplied either by an array of periodic permanent magnets or by an electrically pow- ered solenoid. The requirements for the field range from a few hundred gauss for argon lasers of moderate power operating in the visible portion of the spectrum to more than 1000 gauss for ultra -violet output from krypton or neon. Heat dissipation Most of the electrical energy supplied to the laser discharge tube is transformed into heat which must be removed from the system. Fig. 9 shows a plot of the output power as a function of total input power for several largon-ionlasers. The differencein power in the form of heat is distributed along the bore structure, causing incan- descense of the graphite segments. This heat is removed by radiation through the quartz envelope of the discharge tube to a heat acceptor. Two forms of heat acceptors are currently in use. When the input power to the laser is less than 2000 watts an air-cooled heat acceptor is used. When greater input power levels are required, a water- cooling jacket is used. Mechanical stability Fig. 11-Some of the Argon -ion lasers produced at Lancaster (the size relationships are not Because precise alignment is important accurate). for proper operation of the laser, the entire laser structure, including the dis- Fig. 7. The type of ballast tank em- Table 1-Performance characteristics of RCA lasers charge tube and the mirrors forming ployed depends upon the magnitude of the resonant cavity must be held rigidly the discharge current and the intended Major in place. This stability requires the use duration of operation. Most tubes with Typical wavelengths of rugged cast aluminum mounting gas reservoirs have an operating life of Type No. Gus power (angstroms) structures. The mirror mounts must be up to 10,000 hrs. particularly stable and yet be capable LD2I08 Argon 10 mW 4880 blue-green of precise adjustment to tune the cav- 5145 green Power supply LD2111 Neon 50 mW 3224 Ultra -violet ity. A typical mirror mounts is shown The noble -gas ion lasers typically oper- LD2100 Argon 200 mW 4765 blue in Fig. 10. ate at current levels between 3 and 30 4880 blue-green Commercially available lasers A and with tube voltage drops of 100 5145 green RCA is currently manufacturing noble LD2127Krypton400 mW 3507 Ultra -violet to 300 volts. Fig. 8 shows the voltage - gas ion laser types. Table Ilists the current relationship for a typical opti- 4762 blue 5208 green important performance characteristics mized laser. Because the stability and of each type. Photographs of some noise on the laser beam varies greatly 5682 yellow 6471 red units are shown in Fig. 11. with slight fluctuations in the discharge LD2101Argon 2 W 4765 blue current, the power supplies must be de- References 4880 blue-green 1. Hernqvist,K.G.,"ArgonLasers." RCA signed to have current regulation to 5145 green RCA Reprint RE -15-1-17 within 0.1%. 2. Hernqvist, K. G. & Fendley, J. R. Jr.. "Con- LD2122 Argon 10 W 4880 blue-green struction of Long Life Argon Lasers," /. of 5145 green Quantum Electronics, Vol. OE -3, No. 2 (Feb At the higher power and discharge LD2125Argon 50 mW 1967). current levels, a significant improve- 3. Kozelnik. H., and Li, T., "Laser Beams and (pulsed)4880 blue-green Resonators" /. of Applied Optics, Vol. 5 (Oct ment in power output and efficiency 5145 green '66) p. 50.

59 Design of Cermolox tubes for single-sideband A. Bazarian

Since the conception and design of Cermolox tubes by the design engineering group led by M. B. Shrader at Lancaster in 1956. a large family of tubes has been developed. These tubes cover the frequency spectrum from audio through UHF, and have useful power outputs ranging from a few tens of watts to greater than 15.000 watts. In general. the Cermolox design has provided the flexibility needed for design modifications for many and diverse requirements. This paper investigates some basic problems in the linear tube development and discusses the methods used in the design of these tubes. Albert Bazarian Power Tube Development CERMOLOX TUBES have performed Industrial Tube Division current curves which are parallel and Electronic Components in a wide variety of applications equidistant for equal increments of received the BA in Physics from New York Univer- from communications to radar and grid bias. An operating line, as shown sity and the MS in Physics from the University of counter-measures equipment. In addi- in Fig. Pennsylvania in 1951. He has completed 30 credits la, would be established on of post -graduate studies majoring in gaseous elec- tion, they have been used in applica- these curves by the specification of E,. tronics. His early experience was in mass -spec- tions which required stringent Er I,,, P., and R. The resulting trans- trometer tubes and infrared devices at the Franklin mechanical stability such as vehicular, Institute Laboratories. From 1952 to 1957 he was fer characteristic would be the straight at Chatham Electronics working on basic studies aircraft, and shipboard equipment. The line as shown in Fig. lb. Furthermore, in gaseous discharge and development of gas dis- demands upon the availableRFspec- for a high power -output efficiency, the charge tubes.In1957, he joined the Red Bank trum, particularly by the military, have Division of the Bendix Corporation as a design "knee voltage" would be a small frac- engineer. In 1960, he was promoted to Senior En- necessitated greater use of single -side- tion of the screen voltage. It will be gineer in charge of gas and receiving tube devel- band(ssa)transmission. The high shown that if this tube is operated with opment engineering.His experienceatBendix order of frequency stability (stressed the bias at cutoff (Class B), the result- included development of gas -discharge microwave noise standards, spark gaps, gas lasers,high- by I. P. Magasiny in Ref, 1) and the ingIMDwill be negligible. Class A op- energy xenonflashtubes,and special-purpose reduced spectrum requirements ofSSB eration will also provide negligible iece:vingtubes.He has receivedsix technical permit the assignment of a large num- patents in counting tubes, storage tubes. micro- distortion. Class AB operation, con- wave noise standards, and high-energy transfer ber of channels to the same radio trary to normal expectation,does result spark gaps. In 1965, he joined the RCA power tube spectrum. Cermolox construction in a in considerable levels ofIMD. design group in Lancaster as a product development engineer. He is responsible for power tube ribbon beam formation, without side- Other idealized transfer curves have development programs. Mr. Bazarian is a member rod and lateral wire obstructions by the of IEEE and Sigma Pi Sigma. control grid and screen grid, provides been studied. These curves include quadratic, 3/2 power, and combina- the essential features for the linear where /, is the peak plate current and tubes with the lowest intermodulation tions of curves consisting of a linear upper part and various curvatures of T(e)is an na order Tchebycheff distortion (limn). Present work is con- polynomial. centrated on tubes having 500 W the lower part. Results from these (8791) and 1200 W (8792) of useful curves provide some insight as to how If a two-tone signal such as: power atHFandUHFbands. an actual tube should be designed and how it should be operated to yield the E = A(cos a + cos 19) (2) The single-sideband amplifier must not minimum distortion level. is applied to the grid, it may be shown generate distortion levels above a spe- with evaluation of Tchebycheff poly- cified maximum. With nonlinear dis- Computeranalysis nomials that the resulting plate cur- tortion in the final amplifier, 3rd, 5th. of the transfer characteristic rent is given by 7th and possibly higher orders of inter- \\, hen a signal with varying amplitude modulation distortion (imp) products such as that generated by a two-tone Nomenclature will fall within or near the desired test in theRFspectrum is amplified by sideband. Spectrum crowding requires De plate supply voltage a nonlinear tube, many new frequen- Eb DCplate voltage IMDproducts to be reduced to -45dB cies are generated. The frequency and En,control -grid bias voltage for 3rd, and -50dB for 5th corres- amplitude of theIMDcomponents E.control -grid bias voltage ponding to 0.5 and 0.3 percent, respec- which lie within the fundamental referenced to cut-off tively, of a fundamental tone of a peak fundamental component of bandpass region of the tuned circuits the plate swing two-tone test! may be determined mathematically P.output power with representation of the transfer Reffective plate load resistance Theoretical transfer curves curve by a finite series of Tchebycheff Dc screen -grid voltage A completelylinear tctrode may be E,peak control -grid drive signal polynomials, expanded about a zero En.cut-off voltage characterized as having constant plate- signal operating point: zero signal plate current average plate current h,=Co + C.T.+ C3T2(e)+ +C.T.(e) E,peak Ac component of plate swing Final manuscript received October 8, 1968 (I)

60 Ep

OPERATING LINE IM, puter and a least -square fit method is to 5) the data points of EJE, for vari- 17 EG used to calculate a polynomial ap- ous E. did overlay each other very proximation of the transfer curve from closely. For lower values of E3/E, 5 -EGI the coordinates. The polynomialis however, the location of the cusps

-EGo then used to obtain the expansion shifted for a scan of E. at various bias iG2 El) points for the Tchebycheff analysis. levels. The minimum distortion level PL ATE VOLTAGE -V Tubes have been analyzed in this man- and the number of cusps, however, wj ner and it was found that, generally. always remained the same. drive distortionwillincrease with For the 3/2 -power curve, Class -A oper- signal and will decrease as grid bias ation (Egl E approximately equal to moves toward Class A. 1) would yield low distortion. Other- The 3" and 5" Imo for a typical tube wise the cusps for the 3" and 5" orders -Eco 0 GRID VOLTAGE -V are shown in Fig. 2. The bias is -36.5 do not provide any other convenient (to volts and the signal amplitude ranges low distortion operating point. In ad- Fig.1-Characteristics of an ideallinear from 18.3 volts to 36.5 volts peak. The dition, 5" -order distortion for this tetrode: a) plate current curves; b) dynamic transfer curve obtained from the operating tube output power is also computed curve, as for all others, is generally line for classAB,. and is plotted for the corresponding lower than the 3" except in the regions

te = + h (cos a + cos /3) + ' amplitudes of drive signal. of the cusps where a reversal does

-1-12(cos2a ± cos 2/3) ' occur. /. [cos(2a - /3) + cos (2/3 -a)] + Analysis of ideal transfer curves The distortion curves for the 1.2, 1.6,

+/22 [cos (3a - 2P) + cos (3/3 - 2a) ] + Linear,quadratic, and three -halves and 3.0 -power curves of Fig. 4 are power idealized transfer curves have shown in Figs. 8, 9, and 10. Although + /3,[cos (4a - 3/3) + cos (4/3 - 3a) ] + been computed as shown in Fig. 3. In (3) these transfer curves consist of a + addition, a family of idealized transfer curved and linear portion, the result- where the coefficients 1,1,1 1 are curves consisting of an upper linear ing distortion curves are similar to the polynomials with coefficients Co, C C portion and a lower curved portion 3/2 distortion results. The 3.0 -power

of Eq. I. The various orders of which ranges in curvature from a 3.0 curve does not exhibit any cusps for distortion products are defined as power (cubic) to 1.2 power have been E/E., 10. For this curve, the most follows: studied. These curves range from re- linear operation would occur for Class 3rd order: dB = 20 log (1,11) mote cutoff to rather sharp cutoff A where E,/E is less than 0.6. tubes, as shown in Fig. 4. 5th order: dB = 20 log (1111.) For low -distortion operation, the re- (4) The results of the linear transfer curve sults of the 2.0 -power linear curve of 7th order: dB = 20 log (1111,4) of Fig. 3 are shown in Fig. 5. The very Fig. 11 are promising. For a particular nth order: dB 20 log (A//,k) low distortionat -31.3 volts and bias in Fig. 11, the cusps for the 3 and -62.6 volts bias is representative of 5" do occur at the same ratio for where n is an odd integer and jk is a two digit integer with I = (n - 1) /2 Class -A and Class -B operation, respec- E/E,,. It may be inferred from curve tively. At bias levels of -40.7 volts symmetry that the cusps for 7", 9" and and k = (n + 1) /2. This analytical and -50.1 volts (Class AB,) , the dis- higher orders will likewise occur at the method' provides the basis for a distor- tortion amplitude increases rapidly as same value for El E. This curve is dis- tion analysis of theoretical as well as the signal amplitude exceeds the cut- cussed by Pappenfuss', and, as pointed actualtransfer curves. In addition. off voltage. The computer was pro- out in this reference, if the bias point with the study of the distortion of var- grammed to scan the bias from cutoff is chosen to be at or near the projected ious transfer curves, a technique is ob- to one-half the bias voltage and, for intercept of the linear segment on the tained for predicting the design of each bias voltage, to scan the grid - EGOAVG EGPAA10 actual tube transfer curves and for drive signal from the full bias ampli- °' locating the grid bias on the curve for A tude to one-half amplitude. minimum distortion. This method is 50 500 :k programmed to allow a rapid determi- The quadratic transfer curve of Fig. 3 400 I nation ofIMD.For a given set of input was analyzed and the results are shown data, coefficients of the Tchebysheff in Fig. 6. The slope of the quadratic -100 300 expansion are computed (the C's of was continuous at the point of inter- 200 0 10 20 30 Eq. 1),and then various coefficients of section with the grid bias axis. Itis intermodulation frequencies of Eq. 3 noted that at Class -B operation, the are evaluated.IMDproducts are cal- distortion is high for all levels of grid culated from Eq. 4 by using these signal. As the operation moves toward coefficients. Class A, the distortion decreases and is less for low amplitude signals. In general, when actual tubes are analyzed, measurements of 1. are made at The third and fifth -order distortion points along the operating line at equal products for the 3/2 -power curve of intervals of E, and Et,. These readings Fig. 3 are shown in Fig. 7. It was found are supplied as input data to the corn- that for the higher ratios (EJE.>4 Fig. 2-Typical tube transfer curve analyzed by 61 the Cleary 12 -point system. GRID VOLTAGE- V 25

-10 05

3RD 'pa 30

4 -50 15

0 O

0. -80 a

LINEAR OUADRATIC 05 -110 THREE -HALVES

-130 -100 -90 -60 -40 -20 rai GRID VOLTAGE -V 5 00 Fig. 3-Linear, quadratic, and three -halves F 'E0 power theoretical transfer curves. Fig. 11-Third-and fifth -orderdistortion Fig. 7-Third-and fifth -order distortion prod- products for the 2.0 power curve of Fig. 4. uctsfor the three -halves power curveof CURVE Fig. 3. 1- Y.C1X1 2 2-1,C2X1.6 3-Y*C3X2 -10 4- Y.C4X3

-30

-50

1 EC, g Fig. 12-Causes of curvature of the transfer curve: a) ideal case; b) velocity distribution a of emitted electrons; c) non -uniform electric -100 -90 -60 -40 -20 -90 fieldat the cathode; d) space charge; e) GRID VOLTAGE -V current division between anode and screen Fig. 4-Theoretical transfer curves consist- 1 grid. ing of upper linear portion and various lower -110 curved portions, including curvatures of 1.2, grid -bias axis, then the resultingIMD 1.6, 2.0, and 3.0 powers. -130 products will be very low. In terms of an ideal tube, the most desirable trans- -ISO fer curve would consist of a linear seg- -SO 8 E ment with a sharp slope for highg Fig. 8-Third-and fifth -order distortion prod- and with a zero bias current of suffi- ucts for the 1.2 power curve of Fig. 4. cient magnitude to provide the re- -62 6 quired output power. The lower quadratic portion would comprise only a 31 3 fractionofthetotalcurve length, 20 40 60 80 ,00 thereby insuring a low idling current EG-v Cal and high tube output efficiency.

0, EG 50 Correlation

-50 30 Separate computer programs have been written such that upon specification of the Eb, Er -too E., P., and R, for a a particular tube, the E,. and E. values for the dynamic operating curve for -Co the Tchebycheff expansion points -60 -50 -40 -30 -20 are ECI -V printed out. The transfer curve cor- Isl responding to these points is obtained Fig. 5-Third-order distortion for the linear EG/E0 transfer curve of Fig. 3: a) distortion as a Fig. 9-Third-and fifth -order distortion prod- by measurement, and the curve isana- function of grid drive;b)distortion as a function of grid bias. ucts for the 1.6 power curve of Fig. 4. lyzed for distortion. For a 12 -point analysis, the random errors inmeas- -10 3RD urement of the current may cause a large variation in the computed value -30 5 TX of distortion products. If the transfer

-5 curve is well defined by a large number of readings and by utilization of

-70 the least -square -fit technique, thecor- relation between measured andcom-

990 puted results is considerably improved. The reading error for both lb and E0, 0 -00 have been reduced by use of electroni-

Fig. 6-Third-order distortion for the quad- cally regulated power supplies. With ratic transfer curve of Fig. 3. -130 the use of Tektronix Type -Z amplifier

- i500 Fig. 10-Third- and fifth -order 2 6 distortion products for the EVE, 3.0 power curve of Fig. 4. -3RD ORDER IMD -3RD ORDER MD --5TH. ORDER IMD --5TH ORDER 1610 Ebb2500 V 1v150m4 -35 E02.350 V -35EE2::240500v0v

25 -45 -45 / 250 for recording E, and with a 1% pre- meter upon the static characteristics, . / 350 //:%350 I ----150 IMDproducts, is tested. z ..------/ ii cisioncurrenttransformerforI., as well as the .? -55 _.- -- ,/250 1 .55 ____.....-- ..., z -7 WITH agreement between measured and com- In addition, the effect of any inter- cO --- -7 2 CATHODE ..j, ..... RESISTOR RESISTOR puted values for both 3" and516 IMD actions is detected. Follow-up single - 5 65 ...- / 2: 65 z 1 has been found to be within 2 to 3 dB. parameter tests are used after the effec- 3-35 tiveness of a specific parameter has 350 Tube design been detected. 2-45 -45 it Z 250 The design of a tube having low dis- 2(' ISO Improvements inthe anode design z -.7350 tortion products is based on under- -55 -55 have included reduction of secondary standing of the causes of nonlinearity emission by surface preparation such -65 /7 / 65 AY," of the transfer characteristics. When- 200 300 400 500 as glassblasting and special surface - 200 300 400 500 ever curvatures of the transfer char- PEAK ENVELOPE POWER-W PEAK ENVELOPE POWER-W secondary acteristic occur, the input power is configurationstoentrap Fog. 13-Intermodulation distortion as a function of electrons for improved "knee" charac- power output for RCA 8791 Cermolox tube. distorted in the output circuit. The teristics. Interactions of the anode characteristic features" of actual tubes configuration with optimized screen - Acknowledcments which preclude the construction of an grid -to -anode spacing have been stud- The writer tnanks Mr. D. Mentzer for idealized tube are as follows: ied to establish a minimum level for programming contributions during the 1) Electrons are emitted from the cath- the "knee" potential. conduct of this work. ode with velocity and angular distributions which lead to a distortion Focusing cathodes' have been studied 1. Magasiny,I., "Airborne military transceiver curvature of the transfer curve as shown finds room in crowded spectrum," Electronics primarily for reduction of screen -grid (Apr 15, 1968) pp. 113-138. in Fig. 12, curve D. current. Contoured emitting surfaces 2. Pappenfuss, E. W., Bruene. W. B., and Schoen- 2) The field at the cathode is generally ike. E. 0., Single Sideband Principles and Cir- not uniform. Because of the direct align- are designed to confine the origin of cuits (McGraw Hill 1964) pp. 179-187. the beam to a more limited area. With 3. Cleary, R.E., Approximate Intermodulation ment of control -grid and screen -grid Distorticyi Analysis, Collins Radio Company, wires, the effective emitting area at the the elimination of emitting surfaces Collins Radio Technical Report No. 173 (Oct cathode changes as the electric field is directly behind the control -grid wires, 5, 1956). 4. Pappenfuss, E. W et al, op. cit., pp. 188-189. varied with the grid signal. The result- re- ing distortion curvature is shown in the number of stray electrons is 5. Rodda, S., "Space Charge and Electron De- duced, the effective field at the cathode flection in Beam Tetrode Theory," Electronic curve c. Engineering, Vol. 17, (Jul 1945) pp. 589-592. 3) Space -charge formation at relatively is more homogenous, and the effect of 6. Diemer, G., and jonker. J. L., "Low Distor- high plate currents reduces the effective secondary electrons from the screen tion Power Valves,"Wireless Engineering, Vol. 26 (Dec 1949) pp. 385-389. electric field at the cathode and may grid is greatly reduced. 7. Shade, 0. H. fr., "Beam Power Tube Design cause distortion as shown in curve D. Considerations," Electron Tube Design, Elec- 4) A change in the ratio of screen -grid The control -grid structure of the tube tron Tube Division, RCA, Harrison, New jer- to anode current with a change in has been improved to attain the most sey, (1962) pp. 656-658. 8. Harris,L.A.. Beggs.J.E., and Andrews, anode current produces distortion cur- linear segment of the transfer curve co- C. 1., "Blpotential Cathodes for Reduction of vature similar to curveE.The change Grid Current," IEEE Trans. on Electron De- in this ratio may be due to the follow- incident, with a curved portion which vices, Vol. ED -11, No. 9, (Sep 1964) p. 428. ing causes: yields the lowestIMDlevel. a) When the anode potential swings Table I-Two-tone modulationtestfora below the screen -grid voltage, many Experimental data linear RF power amplifier, class A131 for SSB suppressed carrier service of the electrons deflected by the con- Type8791 has been designed to meet With trol -grid wires cannot reach the an- No 10 -ohm ode and therefore return to the screen specific linearity requirements. Typi- leedbackcathode res. grid. cal test data for two different plate uc plate voltage (v) 2000 25002000 2500 b) Tubes with aligned control -grid voltages with corresponding optimized oc grid -No. 2 and screen -grid wires may show a screen -gridvoltagesare shownin voltage IV) 450350 450350 variation in the number of electrons DC grid -No 1 intercepted by the screen grid as a Table I. For comparison, the effects of voltage IV) -32 -22-38 -21 feedback with a 10 -ohm cathode re- Zero signal result of changes in focusing with plate current (mA) 300300 300300 changes in plate current.' sistor are also included. Effective RF load c) The effect of secondary emission resistance (0) 1850 27501850 2750 Fig. 13 shows the distortion as a func- plate current (mA) at at the screen grid is appreciable and peak envelope may contribute currents greater than tion of peak envelope power output (two tone) 550480 555485 the intercepted portion of the pri- of the 8791 in a two-tone, 30 -MHz, Average plate current mary beam. (mA-single tone) 430 390 430395 linear amplifier with no feedback and grid -No. 2 current An experimental procedure devised to with a 10 -ohm resistor inserted in the (mA) at peak of envelope (two tone)-3.0-4.0-3.5-3.5 study the parameters interfering with cathode circuit. To summarize the tube Average grid -No. 2 the tube design has provided some performance: the third -order intermod- current (mA) -1.8 -2.5 0.4 -4.0 Average grid -No. 1 progress toward the desired tube linear- ulation distortion decreases as the zero. current (mA- ity. This procedure includes systematic signal plate current increases; the ef- single tone) 40 50 45 50 Peak envelope driver/ investigation of the anode design, fect of the 10 -ohm cathode resistor is power (W-approx) 1 1 1 1 focusing cathodes, and design of the to reduce both 3" and 5'" -order dis- Output -circuit efficiency (%-approx) 90 90 90 90 control -grid structure. The experi- tortion by approximately five to seven Distortion product level: mentation technique utilizes an analy- dB; and the 5'6 -order distortion curves 3rd order (dB) -38 -40-43 -44 5th order (dB) -44 -53-50 -51 sis of variance of several parameters at are generally 5 to 10 dB less than the Useful power output 3", and distortion decreases with lower (apprux): more than one level of each parameter. Average (W) 250 250 250 250 In this way, the effect of each para- output powers. Peak envelope (W) 500 500 500 500 63 A sterilizable and ruggedized vidicon Dr. S. A. Ochs F. D. Marschka Fig. 1-Experimental ceramic vidicon. To qualify for a berth on a spacecraft destined to land on one of our sister planets, sisting of ethylene -oxide exposure and a device has to be capable of withstanding pre-flight sterilization as well as the rigors a prolonged 135°C dry -nitrogen bake, of launch and touchdown. The camera tube described in this paper is intended for caused some increase in dark current such an application. It is a one -inch ceramic vidicon with a slow-scan photcconductor and a small gain in sensitivity of the and employs electrostatic focusing and magnetic deflection. The tube is designed to slow -scan photoconductor. The tube produce a high -quality picture after undergoing rigorous test procedures. The sterili- operates satifactorily after undergoing zation requirements presented a potential problem in tube development because of high -amplitude vibration tests, and the possible adverse affect of high temperatures on the photoconductor and the several half -sine shock pulses of 3000- material used for potting the tube within the magnetic shield. The remainingtests con- ' amplitude. cerned the whole vidicon. The tube, shown in Fig. 1, consists of a ceramic body with a quartz faceplate and copper pinch -off; the faceplate is attached to the tube with an indium seal. The filament is potted within the cathode sleeve of the ruggedized electron gun. The wall electrodes consist of metal coatings on the inside wall of the tube envelope, while the mesh MO" decelerating screen is mounted on a .1111 support thatis brazed to the tube body. Metal pins, brazed into the tube wall, provide electrical connections to the various electrodes. The complete tube is potted within a magnetic shield. The ruggedized ceramic vidicon produces as good a picture as the Dr. S. A. Ochs F. D. Marschka, Ldr. more conventional one -inch non -cer- Vidicon Advanced Development Vidicon Advanced Development Industrial Tube Division Industrial Tube Division amic tubes. Electronic Components Electronic Components Lancaster, Pa. Lancaster, Pa. received the BSME in 1943 from Columbia Uni- received the BS in physics and mathematics from Tube construction versity.In 1949 he received the AM and in 1953 Juniata College in 1948, and continued with gradu- the PhD in physics from Columbia. Dr. Ochs was ate work at the University of Marylard He has also The tube design was guided by the an instructor in physics from 1947 to 1951, during pursued graduate studies at Franklin and Marshall following considerations: the course of his graduate studies. Also, while pur- College. Joining RCA in 1949, he has been con- suing his doctorate, he was the recipient of an cerned with production engineering and product 1) All components had to be capable of Atomic Energy Commission pre -doctoral fellowship design engineering of television camera tubes, os- withstanding the required environ- (1951-52). His thesis research involved the use of cilloscopes. and storage tubes. An Engineering mental tests. the atomic and molecular beams technique for the Leader since 1958, he has made significant con- study of the hyperfine structures of rubidium and 2) The cathode surface and photocon- tributionsin advanced development work on the ductor could not be exposed to brazing Potassium isotopes and of sodium chloride.Dr. aforementioned families of tubes. He was instru- Ochs became associated with RCA in 1952, work- mentalin the development work associated with temperatures. ing first on television camera tubes. This included the vidicons used in the TIROS weather satellite. 3) No magnetic material could be used work on a tri-color vidicon, on structure -type vidi- Mr. Marschka is a member of Sigma Pi Sigma. in the middle and front sections of the contargets,and developmental work onnew tube because of the possibility that such image orthicon targets. He later worked on elec- materials might impair picture quality. trostatic signal storage, on tunneling currentin AONE -INCH VIDICON 4) Weight and power requirements insulators,and oninjection -luminescentlasers. capable of Presently, he is engaged in the development of a withstanding sterilizationtreat- were to be minimized. ruggedized, sterilizable vidicon.Dr. Ochs is the ments and severe environmental test- holder of five patents, is represented in the Rein- Fig. 2 is a cross-sectional sketch of the hold Encyclopedia of Electronics with an article ing has been developed for space final tube design. To avoid the weak- on image orthicons, and has received two RCA applications. The ceramic vidicon, est features of the standard vacuum Achievement Awards (1954 and 1957). He is a which represents a significant advance member of the American Physical Society, Sigma tube, the construction method chosen Xi, and a Senior Member of IEEE. in rugged camera -tube design, uses for the ruggedized vidicon consists of electrostatic focusing and magnetic a brazed metal -and -ceramic structure. deflection provided by photo -etched The glass envelope of the ordinary Final manuscript received October 15, 1968. deflection coils to produce an image vacuum tube was eliminated, as well comparable to that of commercial one - The work described in this paper was performed for as the internal glass beads, thin -walled the Jet Propulsion Laboratory under contract inch vidicons using the same photo- electrodes, wire leads, and bulb NAS7-100. conductor. The sterilization tests, con - spacers. The main section of the tube 64 PHOTOCONDUCTOR TARGET LEAD body consists of an alumina -ceramic point); the tension keeps the ring flat FACEPLATE cylinder with internal shoulders for without the need of the usual firing INDIUM SE AL- \ supporting the mesh and the electron operation. As a result, the mesh has a gun. A Kovar section brazed to the relatively high internal damping coeffi- DECELERATING__, rear of this cylinder carries the copper cient that makes it highly resistant to SCREEN METAL COATING exhaust tubulation. damage from severe shock pulses and ON TUBE WALL that minimizes microphonics. The The gun used in the tube is similar in resonant frequency of the mesh ranges geometry to the low -heater -power type from 2300 to 4500 Hz, with most sam- CONTACT PIN used in several commercial vidicons ples resonating near 3600 Hz; the typi- except that its components are brazed cal decay time is 0.2 s. The nichrome together instead of being supported by ring that supports the mesh is screwed CERAMIC TUBE glass beads. As indicated in Fig. 3, the to a molybdenum support ring. The ENVELOPE control and accelerating grids are support ring is brazed to the ceramic brazed to a ceramic spacer which in- tube envelope. sures accurate and stable positioning of the accelerating -grid cup relative The quartz faceplate supports the rho- to the control -grid cup. The control - dium signal electrode and slow -scan grid cup is brazed to a ceramic support photoconductor and is attached to the ring; the ring is brazed to the tube front surface of the ceramic tube body envelope. by a thin layer of indium metal. The indium seal is strong and reliable pro- The cathode heater in the gun consists vided that the ceramic end surface of of a double helix of rhenium -tungsten the tube is polished flat and free of wire just under 0.001 in in diameter; scratches. The electrical connection to the wire is coated with aluminum oxide. the target is made by a nichrome rib- ELECTRON GUN The cathode structure is mounted on bon welded to a stainless -steel ring that the control -grid cup conventionally. encircles the indium seal. The ends of the heater consist of two All other electrical contacts are made straight leads which protrude from the GE TTER long, thin cathode sleeve that contains through the ceramic envelope by mo- ocOvAR SECTION ------the filament. The straight leads are lybdenum or Kovar pins brazed to the COPPER welded to two 0.030 -in Kovar rods tube wall. These pins are connected to TuBULALON whose ends are brazed to the tube wall. metalized strips on the outside surface Fig. 2-Cross section of the new ceramic vidicon. of the envelope. Copper wires are sol- The helical section of the filament is ready to be slipped onto a tube. A yoke potted within the cathode sleeve by dered to the strips near the rear of the consumes about 2 W but could be re- means of loose alumina powder capped tube. The external connections can be designed to require less than 1 W. with a high -temperature cement. seen in Fig. 4 which shows a tube, without deflection yoke, before being The vidicon, with deflection yoke in The filament potting causes relatively potted in the magnetic shield. place, is potted in a polyurethane com- good thermal coupling between heater The vidicon is evacuated through the pound within a magnetic shield of and cathode sleeve. This coupling re- copper tubulation at the rear end. After 0.025 -in Moly Permalloy. The mass sults in a slightly greater heat loss than completion of the standard vacuum - of the total unit is 200 grams. About found in commercial tubes, but per- bake, cathode -activation, and getter - half of this mass is contributed by the mits the filament to operate at a tem- flash procedures, the copper tube is ceramic tube; the other half is con- perature somewhat lower than normal. pinched off to form a vacuum -tight seal. tributed by the potting, magnetic The increase in heater power is less shield, and external wire leads. than 5%. The scanning beam is deflected by the action of a lightweight yoke which fits Environmental testing The electrostatic unipotential lens con- snugly over the tube; the yoke design sists of three cylindrical sections was borrowed from the unit used in The environmental tests outlined in formed by nickel -plated molybdenum the RCA Dev. No. C23080 vidicon. Table 1 were performed by RCA en- coatings on the inside tube wall. The The spiral coils that produce the mag- vironmental engineeringactivityin lens was designed with a magnifica- netic deflection fields consist of photo - Lancaster. Fig. 7 shows the two high - tion near unity with minimum spherical etched copper patterns deposited on a acceleration shock machines designed aberration. Typical operating voltages thin dielectric sheet; each coil con- for this project. The eight -pound mag- are about 400 V for the outside elec- tains forty turns. Fig. 5 is a close-up nesium drop table of the larger ma- trodes and 65 V for the central of two coils; the copper pattern shown chine slides on Teflon bearings and is electrode. is about 0.002 in thick and is depos- capable of providing impact shocks of The decelerating screen consists of an ited on 0.005 -in fiberglass. Two deflec- 3700-g at a drop height of 60 inches. electroplated 1000-line/in nickel mesh. tion yokes are shown in Fig. 6; one has The 0.45 -ms pulse profile for the larger It is stretched over a flat nichrome ring not yet been rolled into the required machine is obtained by using fiber- under high tension (close to the yield cylindrical shape, while the other is glass material as the impact spring. 65 WILL COATIIN DEFINING APERTURE 033 CLCCTROOCI ransom ACCI,TAT PIG

CERAMIC CERAMIC SPACER SUPPORT RING CONTROL GPO vOvAR SUPPORT NEATER ROD CAT.ODE SLEEVE CONTACT PIN

NIGHT 10101TUNE CEMENT RLuM,NA POWDER Fig. 3-Cross section of the gun. Fig. 4-Ceramic vidicon before potting. Fig. 5-Close-up of two photo -etched deflection coils.

The smaller machine was built to pro- to the elevated temperatures required test showed the life expectancy of the vide a 3000-g pulse with a risetime of during the sterilization treatments de- heaters to be at least2000hours. In 0.10 ms; it uses a steel block as the scribed in Table I causes an increase spite of these results, the two complete impact spring. in dark current. However, the increase tubes which were given shock tests of Separate tests were made on all por- is not excessive and is accompanied by 3000 -g'sat the end of the program tions of the tube considered to be an improvement in signal current. In developed open heater connections be- potentially prone to failure; the sepa- general, the dark current grows by 50 fore all sixty shocks had been applied. rate tests were followed by several to 100% while the sensitivity rises In both cases, failure resulted from a made on assembled vidicons. between 15 and 25%. Fig. 8 shows break in one of the straight heater typical tube performance both before sections at or near theKovarsupport Static loads were applied to the face- and after complete sterilization. Dark rods. In the complete tubes, the heaters plate of the tube to establish the current and signal current are plotted were connected to Kovar rods of strength of the quartz disk and of the for a faceplate illumination of one greater length and flexibility than those indium seal. An axial force applied to footcandle;the temperature of the used in the test units. An unexpectedly the faceplate so as to push it away front end of the tube was held at 25°C large deflection of these rods is likely from the tube caused failure equally during the measurements. to have overstressed the straight sec- often from separation of the indium tions of the heaters and caused them seal and from cracking of the face- Sterilization causes a decrease of less to break. plate. Failure occurred with an average than 10% in resolution and an almost applied force of181lbs A lateral imperceptible shift in the spectral re- Static or shock tests were made on force of 60 lbs caused faceplate slip- sponse toward longer wavelengths. No several other components of the tube page. Because the weight of the quartz systematic changes are observed in the including the brazed -in feedthrough faceplateis0.0075lb., a pulse of slope of the transfer characteristic, pins, ceramic-to-Kovar joint, deflection 3000-g causes a peak inertial force gray -scale rendition, or lag character- yoke, and magnetic shield. These were of 3000x0.0075 or 22.5 lbs. This force istics of the photosurfaces. These re- all found to have a considerable im- is well within the limits acceptable to sults indicate that theASOSslow -scan munity to damage when exposed to the faceplate and indium seal and both photoconductor can withstand the ster- 3000-g shock pulses. therefore are judged to be sufficiently ilization procedure without significant The three complete tubes which were rugged for the intended application. impairment of performance. exposed to environmental testing were The tightly stretched nickel decelerat- capable of good performance prior to Photoconductor ing screen was tested under axial shock the shock tests. One tube was passed The slow -scanASOS(antimony sulfide- in a test arrangement designed to simu- through the entire ethylene -oxide de- oxysulfide) photoconductor, developed late the mounting used in the vidicon. contamination and heat sterilization by RCA for earlier space applications, A considerable number of samples was procedures, as well as the static accel- was chosen for this tube because pre- exposed to shocks of 3000 -g's for 0.45 eration, sinusoidal vibration, and wide- liminary tests had shown it to be rela- ms; no permanent damage or wrin- band noise tests, without deterioration tively resistant to high temperatures. kling occurred. in tube performance. However, the fila- The photosurface is deposited on the ment opened after one 7500-g shock faceplate; a substrate of vitreous A major part of the design effort was test. The other two tubes were exposed quartz of very high surface quality. spent on the brazed gun structure to to shock tests in which the amplitudes assure satisfaction of the two principal To avoid coherent noise caused by im- were held to slightly above 3000-g. perfections on or in the photoconduc- gun requirements: precise alignment One tube was in good condition after and strength. The final gun design tor, the faceplate is subjected to a very the first set of five shocks, but devel- shown in Fig. 3 provides the necessary thorough cleaning process. The signal electrode and the photoconductor are accuracy and has been shown to be Table I-Environmental tests performed on capable of withstanding shock pulses the vidicon evaporated in an oil -free system evacu- of 3000 -g's. Ethylene -oxide decontamination: six 28 -hr cycles ated by absorption and ion pumps. All at 50°C. Dry -heat sterilization: six 92 -hr cycles at 135°C. depositions are made in a single pump- Several potted heaters were mounted Static acceleration: six tests at ±190 g for 20 min along three orthogonal axes. down-i.e., without breaking vacuum in special test units and exposed to the Sinusoidal vibration: tests of each type along three between the evaporation of different ma- complete set of 3000-g shock pulses orthogonal axes with vibration swept at1/2 oc- tave/minute. up and down in frequency. terials. The continous vacuum assures described in Table I; all of the samples ±0.5 in displacement, 5 to 16 Hz. a sensitive and spot -free photosurface. survived. An accelerated life test run 5 g RMS. 17 to 50 Hz. 15 g RMS, 50 to 100 Hz. As expected, the prolonged exposure on the samples following the shock 35 g RMS. 100 to 2000 Hz. Wideband noise: three tests at 25 g RMS, 9 min duration. 15 to 2000 Hz along each axis. Shock: sixty tests with designated shocks applied 5 times in each of 6 directions. ±3000 g half -sine pulse, 225 As risetime. -±-3000 g half -sine pulse, 100 gs risetime. /

Fig. 6-Two photo -etched deflection yokes. T ORGET vOLTAGE -V Fig. 8-High-acceleration shock machines. E zrosuat--rooTosedi-scoNos oped an open heater connection during plete erasure can be obtained by scan- Fig. 9-Transfer characteristic. the second set. The second tube went ning the photosurface several times, through twenty-five shocks with no possibly at a fast rate, during the time damage but lost heater continuity dur- alloted to the erase step. ing the following five shocks. The failure of each of the three tubes was The transfer characteristic shown in caused by breaks in the straight sec- Fig. 9 describes the sensitivity of the tion of the heater wire. tube in terms of the signal current gen- so erated for various faceplate exposures.

The tube was operated in a two -second gt 70 Tube performance cycle with one second used for the readout. A reciprocal relationship was The performance characteristics of the 6, ceramic vidicon are described in Figs. found to exist between shutter time 9, 10 and 11. In slow -scan operation, and illumination (at least within the Isec- a typical cycle consists of three steps: shutter -time range of 1/100 to 50 exposure, readout, and erase. In expos- ond). Therefore, for a given exposure 0 ure, the lens shutter is opened for a value(the product of illumination 40 specified time. The photosurface then and shutter time) the signal remains carries a charge pattern corresponding the same if, for example, the illumina- to the scene whose image appears on tion is doubled and the shutter time is 30 the tube. In readout, the electron beam halved. The transfer characteristic, in scans the photosurface and generates logarithmic form, is essentially a the video signal in accordance with straight line with a slope y of 0.71. the charge stored on the photosurface. This transfer characteristic is typical During the erase step, the electron of vidicons and permits operation over beam scans the photosurface a second a relatively wide range of exposure time and removes most of the remain- times. )) 200 300 400 500 700 ing stored charge. The erase scan is The resolution capability of the tube TELEVISION LINE NUMBER desirable because the beam ordinarily is shown in Fig. 10. The uncompen- Fig. 10-Resolution capability. removes only part of the stored charge sated horizontal peak -to -peak reponse in any single scan. In a typical experi- at the center of the raster is shown for ment, the beam erased slightly less a square -wave test pattern with the than half of the remaining stored tube operated at the standard televi- charge with each scan. Therefore, after sion rate of 30 frames/second. The the readout and erase scans, about performance is essentially the same as two-thirds of the originally stored that of an RCA -8134 vidicon, a com- charge had been removed. More commercial one -inch tube with electrostatic focusing and magnetic deflection. Fig. 7-Tube performance before and after sterilization (standard frame rate). Fig. 11 shows the spectral response of the ruggedized tube in the visible -light range; power input at all wavelengths is the same. As indicated in the figure, the photoconductor has a maximum sensitivity at some wavelength between 500 and 600 nanometers. 2

Acknowledgements The authors acknowledge contributions to this work by J. Ziedonis in environmental engineering, L. Rhoads 0 400 500 600 in ceramic engineering, and D. Neuer WAVELENGTH - nrn in tube construction. F.g. 11-Spectral response.

67 power at the cable surge impedance Broadband, high -gain, L -band, (usually 50 ohms) and thus makes interstage performance independent of cable length. Interstage loading and power amplifier module tuning adjustments are made with an R. L. BaileyJ. R. Jasinski impedance transformer at the generator end of each cable. The overall gain Increasingly sophisticated radar systems. such as phased arrays, have creatada need of the amplifier chain increases with for RF power amplifiers that have wide bandwidths together with excellent phase the addition of each new single -tuned stability and uniformity of characteristics. This paper describes the unique dasign and stage, while system bandwidth con- fabrication of the RCA Dev. No. Y1043', a six -stage. L -band amplifier module that has tinually decreases;desired gain - performance characteristics which fulfill these requirements. bandwidth characteristics may be achieved by changes in the degree of coupling between stages and/or by the AMPLII II. RSutilizing grid -controlled consistof tuned input and output tubes have been operated suc- cavity or stripline resonators in con- use of stagger -tuning. cessfully innarrow -band phased -array junction with the appropriate bypass Fig. 1 shows a single -tuned, single - radars for a number of years; only circuits required for operation. A typi- stage amplifier, RCA Dev. No. Y1015. recently, however, have their full cal 5 -kW unit may be 5 to 7 inches on This amplifier uses an RCA -7651 tube, capabilities inbroadband,high -gain each side in cross section and 12 to 15 delivers a minimum peak power out- applications been demonstrated. The inches in length. These amplifiers are put of 6 kW, and can be mechanically development of the Y1043 amplifier normally single -tuned, and yield char- tuned across the 950 -to -1225 -MHz band module has substantially advanced the acteristic gains of 10 to 15 dB. Gain is with an average instantaneous band- state-of-the-art in this area. The com- increased when two or more inde- width of 3%. The RCA -7651 is capable pact, wide -band, pulsed amplifier mod- pendent stages are cascaded with of a peak power output of 39 kW at ule uses gridded electron tubes, and is interconnectingcoaxialcables.An 1200 MHz when operated at its maxi- designed for operation at a center fre- impedance -matching transformer at the mum ratings in a narrow -band circuit quency of 1300 MHz with an instan- load end of each interconnecting cable (less than 1% instantaneous band- taneous electronic bandwidth of 10% permits transmission of interstage width). The threaded coaxial fittings measured at the 1 -dB points.' Perform- ance objectives included a peak power J. R. JasInski R. L. Bailey output of 5 kW and an overall gain of Regular Power Devices Engineering Advanced Development, Power Devices Industrial Tube Division Industrial Tube Division 45 dB utilizing a minimum number of Lancaster, Pa. Lancaster, Pa. stages. The amplifier module was lim- was an honor graduate of theT-3 Engineering received the BSEE in 1958 from Washington State ited to 24 inches in length and 41/2 by curriculum, RCAInstitutesin1958.Hejoined University, and the MS inPhysics in1968 from RCA in 1958 as an electronics technician at the Franklin and Marshall College. He joined RCA in 41/2 inches in cross section. In this par- Lancaster plant, where he was engagedinthe 1958 asa design engineerinthePower Tube ticular phased -array application, each evaluation of RCA's new line of Cermolox tetrodes department, where he worked until 1963 on the to establish their capability in pulsed RF service design and development of UHF circuits.In 1963 amplifier module was to drive an indi- atL -band frequencies.In 1962, Mr. Jasieski was he joined the Power Tube Advanced Development vidual radiating element. For proper assignedtoevaluationoftheS -band coaxitron group, where he was engaged until1965 inthe antenna spacing, this modular approach being developed for the AirForce. Compilation design of a broadband, high -gain L -band ampli- and organization of data from this evaluation led fier. Since 1965 he has been engaged in develop- required that the amplifier not exceed directly to the successful development of external ing techniques for producing high power from RF one half wavelength in cross section. broadband circuits for Cermolox tubes designed to transistors. Mr. Bailey isa member of IEEE and operate at L -band frequencies. In 1964. Mr. Jasin- Sigma Pi Sigma. skiwas assignedtothe developmentalL -band The degree of phase variation from in- module program. He was upgraded to Engineer put to output as a function of drive in 1967 and is presently responsible for the proto- The developmentofthegridded-tubeL -band, level, frequency, and supply -voltage type development of all new module variants and power -amplifier module was supported by the manufacture and testingofexisting module U.S A.E.L.,EvansLaboratories,underContract variation was required to be small to types. DA36-039-AMC-03203(E). minimize the cost of auxiliary operating equipment.

Conventional amplifiers Analysis of the specifications described above indicated that conventional cascading techniques would not yield the desiredresults,particularlyifthe overall size and gain -bandwidth requirements were to be achieved. Con- ventional L -band (390 -to -1550 -MHz) gridded power -tube amplifiers typically

Final manuscript received March 13, 1969.

68 uses two commercial 7768 planar tri- input sections of the tubes; as a result, odes (small tubes) and four cermolox the amplifier stages have an average tetrodes(two each 8226 and 7651 length of only 4 inches. Two over - generic types) , samples of which are coupling elements are contained in the shown adjacent to the module. The interstage circuit external to the tube cermolox tetrodes used in this ampli- (about one-half wavelength) to pro- fier differ from other tubes of the same vide optimum gain -bandwidth ina generic type in that they use integral minimum amount of space. In each conduction -cooling aluminum jackets stage, these elements consist of shunt instead of air-cooled radiators. TheRF inductance and series of capacitance. drive, liquid coolant (high purity at 1/2 The inductor is composed of several GPM) , and all operating voltages are 1/4 inch -diameter spokes shunted across applied through the five connectors at the cavity. In each tetrode stage, two of the input bulkhead. Output power of the spokes are hollow to allow anode the amplifier is extracted through a coolant to flow into and through the coaxial connector at the opposite end. anode seat, which is a portion of the A 3 -kVDCpower supply, the highest coaxial center conductor. The series voltage required for operation, pro- overcoupling capacitor is formed by a vides plate voltage for the four tet- break in the center conductor of the Fig. 1-Conventional Y1015 single -tuned, single -stage amplifier rodes. The tetrodes are modulated by interstage coaxial line. This capacitor shown in Fig. I permit connection of a 1 -kV pulse which is applied to the also isolates the plate voltage of a par- screen grids. The plate voltage for the ticular stage from the cathode poten- standard 50 -ohmRFdrive andRFout- put cables to the cavity; however, 7768 triodes is provided by a 300 -volt tial of a driven stage. The module RFradial bypass capaci- because the cables join the coaxial DCsupply; the grids are modulated by contains mica tors which provideDCisolation for the resonators at right angles to the main a small 6 -volt bias pulser. All voltages cavity axis, they add considerably to are applied through bus lines which various tube elements. thecross-sectionaldimension.This are located in the corners of the module. A few of these bus lines are Basic design techniques cavity measures about 6 inches at its Overcoupled RF circuits major cross section. When the cavity visible in Fig. 2. The module is pack- TheRFcircuits used throughout the is tuned to its lowest operating fre- aged in an aluminum container which measures 43/8 x 43/8 x 23 inches. This module are overcoupled resonant co- quency of 950 MHz, the input tuning axial cavities which are aligned to rods extend 2 inches below the cavity size satisfies the required dimensions for use in a 1300 -MHz phased -array produce triple -tuned, maximally flat and produce an overall length of 141/2 responses. Each stage in the chain has radar. inches. the same response. This approach pro- This type of amplifierisgenerally Fig. 3 shows the instantaneous band- vided full freedom to change the num- not satisfactory for use in multistage width characteristic of the module for ber of stages at any time during the broadband chains. Because the unnec- an applied peak drive power of 25 development of the module. The alter- essary impedance transformations to mW. The module measured produced native approach, stagger tuning of indi- the 50 -ohm interconnecting cable surge a nominal peak power output of 5.75 vidual stages, would have restricted impedance severelylimitsgain - kW at 2% duty, with a ripple factor of this freedom. bandwidth product, double- or triple - less than 1 dB and an instantaneous The overcoupled circuits have an tuned circuits cannot be utilized to bandwidth of 10.5%. inherently greater gain -bandwidth best advantage, and the resulting size Fig. 4 shoals the final amplifier pack- product than the simple single -tuned of the cascaded amplifier chain exceeds age. The absence of connectors on the circuit; the improvement is a function the requirements of many applications. large surfaces allows many such mod- of the number of circuits coupled to- ules to be inserted side -by -side, "honey- gether. A useful figure of merit for New cascading technique comb" fashion, into an array panel to broadband circuits is the resistance - Fig. 2 shows the RCA Y1043 "totem - form a multi -megawatt phase-steerable times -bandwidth (R x BW) product pole" amplifier. This amplifier incorpo- beam. Fig. 5 shows a cross-sectional which the circuit presents to the funda- rates techniques which enhance the view. No conventional, lengthy inter - mental component of the tube beam inherent gain -bandwidth capabilities stage circuits are used in the amplifier current. A circuit can be evaluated on of gridded tubes and, at the same time, chain; instead, the stages are stacked the basis of the ratio of the actual achieves a considerable size reduction in a "totem -pole" fashion, and each of R x BW product to the maximum over classical cascaded amplifier chains. the stagesisdirectly interconnected R x BW product for a simple parallel The amplifier is shown with the exter- by a full -electrical -wavelength circuit. resonant circuit having the same ampli- nal housing removed to illustrate the Minimum system dimensions are pro- fier output capacitance. A valid com- end -to -endconfigurationasviewed duced with this coaxial, in -line con- parison, however, must include a from the input end. In this system, six figuration. A considerable portion of specification of the shape of the actual stages of gridded power tubes provide each interstage circuit, especially in response. For thisdiscussion,itis a peak power output of 5 kW at 1300 the tetrode stages, is contained within assumed that the response is maximally MHz with a gain of 53 dB. The system the internal output and succeeeding flat and, further, that the R x BW value 69 is the product of the resistance and 7651 cermolox tetrode was chosen for bandwidth both referred to the points the final stage; an equivalent circuit 1/6 dB down from the peak value of R. for this stage is shown at the top of (Because the module has six stages, Fig. 6. The impedance and admittance thisreference providestheoverall plots for such a circuit can be obtained R x BW product at the -1 -dB points.) from straightforward but tedious cal- For the single -tuned circuit, the R x BW culations. A program was written for product is the half -power resistance the RCA -301 digital computer to times the half -power bandwidth. greatly improve the calculation time. The program included provision for 15 On this basis, an optimized double - different sections of transmission line, tuned circuit can provide an R x BW and made allowances for lumped re- product as great as 1.22 times that of a active elements at various line junc- single -tuned circuit. A triple -tuned cir- tions. The output data provided the cuit can provide up to 1.66 times as impedance as a function of frequency much R x BW product. The limit ap- at key line junctions. Fig. 2-Y1043 amplifier module with cover removed to proaches 7r as the number of tuned show the end -to -end configuration. circuits is increased. (These numbers The impedance and resistance plots in were calculated from tabulated data Fig. 6 are comprehensive presentations 7.0 for ladder networks given by Wein- of data obtained from computer analy- 7".\\,...... /-N berg': verification of the extension to sis of the conceptual design of the final 60 high -frequency bandpass circuits was output stage. The particular values list- %.0...... " \ 5.0 made by Green.') Because the circuit ed for the circuit parameters produced / complexity increases directly with the theSmithplotvariationinim- 4.0 \ number of coupled circuits, a triple - pedance Z,Z0 as a function of fre- 1 3.0 tuned circuit provides a good compro- quency shown at 0 PULSEWIDTH1000s the lower right. DUTYCYCLE 2% mise between R x BW improvement Within the passband, the impedance RFDR1VE.0.025WATT 2.0 II 4 and circuit complexity. response follows a constant R locus, SAND /' OF indicating a maximally flat condition. 1.0 INTEREST The triple -tuned coaxial circuits em- The resistive component of Z, is ployed in the Y1043 L -band module 0 SO 00 so plotted at the lower left. Obviously, 1200 20 SO 1300 20 '0 1400 are very similar conceptually to the FREQUENCY-NM the first set of parameters used in the double -tunedcircuitsemployedin Fig. 3-Y1043 typical output response. RCA super -power coaxitrons."." computer calculation did not produce this optimized response, because the values of the parameters were chosen Tub* selection and computer calculations primarily by educated guesses. How- Because the R x BW product for the ever, with the Smith plots used as simple parallel resonant circuit can be guides, successive computed sets of easily calculated for a hypothetical data converged rapidly. lumped -constant circuit, it is possible to estimate the performance of any During the series of computer calcula- tube used with triple -tuned circuits. tions to determine the specific set of Tubes for use in the amplifier module parameter values for the optimum were selected on this basis. For inter - response, much valuable insight was stage analysis, the dynamic drive -point gained on quantitative effects of vary- impedance was measured at the input ing each parameter. This information terminals of each tube at the required was later found to be very useful for drive power. A modified version of the circuit alignment. The computer re - Fig. 4-Y1043 shown fully housed to illustrate Its modular concept. Fig. 5-Tube cross section. 9, 93 FI

. 11..P701111.1 ,amene.. c77-- Ell 11 ar

SHUNT SEIZIES SERIES SERIES 02 1 SHUNT SHUNT TUNING IAN= COOLANT) (ANODE COOLANT)

Trie- A211133 - - 70 - '4" ------AZIP211 SUS-ASSENSLY SUS -ASMARA' IRAI-ASSENSLY Each subassembly contains two stages ages, the filament voltage, and drive

s Lif%0LOAD.O DuCTOO of amplification, and each was exten- power produced negligible phase OCP0- MEW CAPACITANCE sively tested as a single -stage and then changes through the module. SO a two -stage broadband amplifier to con- 340 SOD The phase characteristics as a function COOPLIPG 040,P TOP firm successful operation of each tube 45. OuMIT GaP of frequency, shown in Fig. 8, indicate ALL FROIAPLy- DIP( Wolfer ...LES at the required power level. The input lel ant Pt xlEPIPCIO TO aoo to. that extreme linearity(±2°)exists P, X M. 20.500 0.4.- Moo to the first stage was not triple -tuned, WPM 11, RAP 5(005 over 75% of the pass band with no CPI. MP. POT.. THE I/G- CH POWS but used a shunt capacitive "matching significant irregularities appearing in slug" that produced asingle -tuned the curve. ZOO transformation to 50 ohms. Because the transformationratiowasrelatively Extendedapplications

To low, the input%SWRisbetter than Since the successful prototype develop- 1.5: 1 (4% reflected power) across the 0 I ment of the Y1043 module, customer us° Sao aso p350 1200 -to -1400 -MHz band. TRIOUIPC- interest has warranted the establish- Final assembly and alignment pro- ment of a manufacturing facility for ceeded progressively, one stage at a fabrication of such modules on a con- Fig. 6-Smith-chart plot of impedance versus frequency. time, starting with the low -power tinuing basis. Although the module was stages. With this method, two stages primarily designed for phased -array sults were used to calculate projected were operated into a passive load, then applications, present interest includes operating conditions for each stage. In three stages and finally up to six stages other areas where small size and high practice, it is necessary to derate the so that alignment of the most recently gain -bandwidth requirements are also computed value of R, by approxi- added stage could be accomplished at of prime concern. The Y1043 module, mately a factor of 30% to compensate the correct drive level. TheDCgrid - and variants thereof, have either been for the practical limits on circuit effi- current response of the driven stage proposed for or are presently being ciency and for the smearing of theRF served as an excellent indicator of used inRFpulse -coded systems for current pulse because of transit -time interstage performance. The use of command and guidance, phase -stable effects. swept -frequency drive at all times pro- broadband drivers for higher -power vided comparative ease of circuit ad- radar systems, multi -channel airborne Cold -probe verification of amplifier design justment. navigational beacons, and broadband After paper designs of circuits were applications where the major interest available, cold -probe models were fab- Phase measurements is in frequency agility. ricated and analyzed to verify predic- The phase characteristics of the module As an illustration, the RCA Y1048C tions based on computer calculations. are of major interest for use in phased - These models consisted of cavities that amplifier shown in Fig. 9 was designed array systems. Two independent meas- specifically to operate in one of the included provisions for variation of urements of these characteristics were every parameter external to the tube; above application areas and is an ex- made with nearly identical results, one tensionof Y1043 techniques. The they used dummy tubes consisting of at MIT Lincoln Laboratories and the only screen -grid -and -anode assemblies. module consists of 5 cascaded tetrode other at RCA Lancaster. These meas- stages. It is designed to operate at a The output gap of the dummy tube was urements were made with a differential center frequency of 1375 MHz with an energized with approximately 1 watt of phase bridge which had a resolution instantaneous 1 -dB bandwidth of 5%. frequency -swept power, and a crystal capability of better than three degrees. The peak power output is 10 kW with detector was used to monitor the trans- The measured phase -sensitivity data a gain of 47 dB. Because this module mission response of the cavity into a are presented in Fig. 7. As expected of was not to be used in a phased array, 50 -ohm load. The output of the crystal any device that utilizes gridded power the cross-sectional dimensions were detector was then displayed as an os- tubes, phase sensitivity of the module allowed to increase to 5x6 inches; the cillogram.Variationofparameters is extremely low. The most sensitive lengthis 22 inches. The increased while the instantaneous display was operating parameter, the tetrode cross-sectional area was required for monitored allowed swift attainment of screen -grid voltage, produced an ap- secure mounting of the module and its a flat response with the desired center proximate shift of only one degree per auxiliary components within the alu- frequency and bandwidth. Results ob- 1% change in voltage. As a result of minum container to meet shock and vi- tained with models of the final output this high degree of phase stability, pow- bration specifications. and the interstage circuits verified that er -supply costs are reduced because the computer program accurately sim- voltage -regulation requirements are RCA Dev. Nos. Y1057 and Y1049 modules (not shown) are the one- and ulated theRFcircuitry. minimized. As a direct comparison, a velocity -modulated device, such as a two -kilowatt versions of the Y1043. Amplifier assembly and alignment traveling -wave tube typically exhibits a The Y1057 delivers1 kW of peak The final design and fabrication of all twenty -degree phase change for a 1% power at a center frequency of 1270 the amplifier stages were based com- change in anode voltage"; these results MHz with an instantaneous 1 -dB band- pletely on computer and cold -probe re- show that much better voltage regula- width of 70 MHz and a gaing of 37 dB. sults. As shown in Fig. 5, the module tion is required for the same degree of The Y1049 delivers 2 kW of peak consists of three basic subassemblies. phase stability. Variations ofDCvolt- power at a center frequency of 1088

71 Parameter varied 1235 MHz 1265 MHz 1300 MHz 1335 MHz 1385 MHz I. 3,000-V supply SPECIFIED FREQUENCY RANGE -Y. (tetrode plate) 20 Lowered 10% + .3 - 4.25 Lowered 16.7% - .45 30 iv II. 1,000-V supply (tetrode screen) so Lowered 10% + 6.7 + 55 + 7.2 + 9.75 +11.1 Lowered 20% +169 +135 +15.6 +24.8 +23 6 o III. 300-V supply (triode plate) ao ....- Lowered 10% APPROXIMATE W + 3.76 4--FORa rUNEARITY4 IV. Filament supply 90 194 NFU 7-1/470 (all stages) Lowered 8.75% - 4.38 1200 1250 1300 1370 1400 V. Drive power FREQUENCY -MN/ Lowered 2dB - 2.8 Lowered 4dB - 7.2 Fig. 8 -Measured shiftin phase with fre- VI. Drive Lowered 1dB 3,000-V, quency for the module. 1,000-V, and 300-V supplies Lowered 10% 14. 12.7 4.9 be arranged in a very compact circuit. Fig. 7 -Phase sensitivity data The design and analysis of these complex circuits are facilitated by the use MHz with an instantaneous 1 -dB band- MHz, respectively. Thus, the 1235 -to - of a digital computer. Several stages of width of 70 MHz and a gain of 37 dB. 1365 -MHz operating spectrum for the power tubes with maximally flat re- Both modules are capable of operating Y1043 permits use of the ideal circuit at 5% duty and each contains a single configuration. The larger tube types, sponses can be cascaded or stacked tube type, the RCA Dev. No. A2833; because of their greater power capabil- "totem -pole" fashion to form a com- the Y1057 has four stages and the ity, have inherently lower strap reso- pact, high -gain amplifier module. Re- cent circuit refinements have produced Y1049 only three. Proposals have re- nant frequencies. The upper frequency cently been written for design and fab- limit is approximately 600 MHz for the overall gains of 60 dB, corresponding rication of 50k W and 100k W modules 100kW module and about E 90 MHz to 2.6 dB per linear inch of module. to cover the 750-to850 MHz and the for the 50kW module. Measurements on a six -stage module 405 -to -450 MHz bands, respectively. Cross-sectional size and totLl length have proven that the excellent phase All of the above circuits use the same primarily determine the lowest prac- stability of gridded power tubes is triple -tuned, coaxial, "totem -pole" cir- tical operating frequency. Lcwer fre- preserved through as many as six stages cuit concepts described for the Y1043 quencies require an increase in cross ofamplification. The phase -versus - frequency characteristics of the module module. For a particular tube type, the section for maintenance of effectiveRF are such that phase -linearity compensa- high -frequency limit is determined by radial bypassing of the mica roc block- tion can be accomplished easily. the internal screen -grid -to -anode "strap ers. A typical 800 -MHz module would resonance" of the tube, i.e., the fre- have to be increased to 8x8 itches in quency at which the shunt -over -coup- cross section to prevent blockEr radia- Acknowledgments ling inductor appears just at the tube tion. Another consideration fo- lower - The authors thank W. P. Bennett, F. W. terminals. Attempts at circuiting at fre- frequency operation is water-cooling Peterson, M. V. Hoover, and L. F. quencies above "strap resonance" requirements. As the operatioial fre- Heckman for their important contribu- would produce a loss of this shunt - quency is decreased, the shunt induc- tions to the design and testing of the inductor "tuning handle". For a given tors move further from the tube and module described; H. Kaunzinger and frequency, the ideal circuit has the anode seat; hence, a more c pmplex G. Fincke of USAEL, the technical shuntovercouplinginductorsvery mechanical configuration is required monitors of the contract work, for their close to the output terminals of the for cooling. many helpful suggestions; and L. Cart- tube. This condition produces inter - ledge and others of. MIT Lincoln Lab- stage circuits of the shortest possible Conclusions oratoriesfor measurementsofthe length and optimizes the dimensions of phase characteristics of the module. themodule. The upper frequency High -gain, broadband amplifiers using limits of the tetrodes used in the Y1043 triple -tuned coaxial resonators in com- References are approximately 1410 MHz and 1600 bination with gridded power tubes can I. Bailey,B.L.,"A Broadband,High -Gain Power AmplifierforPhased Arrays Using Phase -Stable Gridded Tubes," IEEE Interna- tional Convention, New York, N.Y. (Mar. 26 1965). 2. Weinberg, L., "Optimum Ladder Networks" Journal of Franklin Institute (July 1957). 3. Green, E., "Amplitude -Frequency Character- istics of Ladder Networks," Marconi's Wire- less Telegraph Co., Ltd. 4. Parker, W. N., "The Coaxitron Tube," RCA Pioneer (Feb 1961). 5. "The Development of an Experimental Model of a High -Power, Broadband Coaxitron RADC-TR-60-94 (June 1960). 6. "The Continuaiton of the Development of an Experimental Model of a High Power Broad- band Coaxitron Amplifier," RADC-TDR-61- 315 (Nov 1961). 7. Keith, F.S., Parker, W. N., Rintz, C. L., "The Coaxitron-A High -Power, Broadband, Integral -Cavity UHF Amplifier," AIEE Gen- eral Meeting in New York (Jan 31 1962). 8. Molz, K. F., "Selecting Tubes and Receivers for Large Phased -Arrays" Microwaves (Mar 1964) p. 18.

72 Fig. 9-Y1048C five -stage ruggedized module with side cover removed,

41r