- / ■^?5 >i':\>y ..<:ublication Section, Technical Information Branch, Atomic Energy Commission, P. O. Box E, Oak Ridge, Tennessee. Inasmuch as a declassified document may differ materially from the original classified document by reason of deletions necessary to accomplish declassification, this copy does not constitute authority for declassification of classified copies of a similar document which may bear the same title and authors. Date of Manuscript: July 10, 1946 Document Declassified: May 19, 1947 This document consists of 14 pages. MDDC-983 Table of Contents Page 1. Purpose 2 2. Introduction 2 3. Principles of Operation 2 4. Magnet - Structure 3 4.1 Coils and Cooling 3 4.2 Excitation and Control 3 4.3 Other Arrangements 4 4.4 Shims 4 5. Vacuum Chamber - Structure 4 5.1 Vacuum Seals 5 5.2 Vacuum I>umps 6 5.3 Lining 6 6. Accelerating Electrode - Structure 6 6.1 Adjustment ^ 6.2 Support 7 7. Ion Source 9 8. Resonant System - General Requirements 9 8.1 Frequency Modulation 10 9. Ion Deflector Systems 11 10. Target Arrangements 11 11. Shielding 12 12. Control Circuits 12 Digitized by the Internet Arclnive in 2011 witli funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation http://www.archive.org/details/universityofcaliOOusat -1- MDDC-983 ABSTRACT For declassification purposes, a general description of the California cycloti"ons is given. The discussion is on an elementary level and indicates the scope and function of the various components described in detail in a large mass of engineering drawings and special reports. -2- . MDDC-983 THE UNIVERSITY OF CALIFORNIA SYNCHRO-CYCLOTRONS By R, L. Thornton 1. Purpose The intention of this report is to describe in general terms the 37" and 184" synchro- cyclotrons developed in recent months at the Radiation Laboratory. This report will then be reviewed for declassification in the usual manner, and it is our understanding that upon declassification, detailed blueprints and reports covering the material disclosed in this re- port may be forwarded to seriously interested laboratories (in this country or abroad) to assist them in their accelerator programs, at our discretion and without further declassi- fication. It is also our understanding that the 60" cyclotron of the Crocker Laboratory which was completed in 1939 is not regarded as classified. Such modifications and improvements as are made on this equipment from time to time parallel in general the equipment described in this report and declassification of such modifications is also requested by this report. 2. Introduction The conventional or constant frequency cyclotron as devised by Prof. E . O. Lawrence and subsequently developed here and elsewhere has been adequately described in scientific Journals. Two review articles by M. Stanley Livingston appeared in the Journal of Applied Physics (15, 1, 1944 and 15, 128, 1944); these articles, together with the references to earlier publications cited therein, constitute an adequate description of such cyclotrons. The state of the art as covered by these papers will be assumed as well known,and this report will confine itself in general to a description of the modifications and changes made here on the 37" and 184" machine. Since these two machines are in most respects similar, major emphasis will be placed upon the 184" machine and reference made as required to the differences existing between them. A s3mchro- cyclotron for Harvard University is under design here and similar ref- erence will be made to special design features of this 92" cyclotron as well as to a some- what larger machine under consideration for the University of Rochester. 3. Principles of Operation It has been shown theoretically that, independent of the time an ion is drawn from the ion source, it gets into phase with the radio frequency voltage, i.e., the ion crosses the center of the gap at peak dee voltage, within a few revolutions. Since in a constant frequency cyclo- tron,the ions will gradually fall behind in phase due to the relativistic increase In mass and the radial magnetic field decrease required for focussing, it is clear that after a sufficient number of revolutions the ions will be out of phase with the voltage and further acceleration will become impossible. It is also clear that the higher the dee voltage and hence, the greater the energy gain per turn, the greater will be the final energy reached when this condition occurs. This situation has been investigated in detail by Rose, Wilson, and others. Clearly, however, If the frequency of the dee oscillations decreases at the proper rate, -3- MDDC-983 then the ions will remain in phase and this limit to the attainable energy will not exist. It is the principal contribution of the theory of phase stable orbits developed by Vekeler, and independently by McMillan, to point out that, providing the ions cross the accelerating gap in the part of the cycle when dee voltage is increasing with time, the phase of the ions will oscillate about an equilibrium phase determined by the gain in energy per turn required to keep in step with the changing frequency. Thus, it is not necessary to hold the frequency- time curve accurately to a predetermined shape, but tlie ions will automatically adjust them- selves to the conditions existing within wide limits, provided obviously that the required energy gain per turn does not exceed that available from the peak dee voltage applied. De- tailed examination of the ion orbits by McMillan,Bohm^and others has shown that the neces- sary damping on the various possible oscillations exists and that the starting conditions are such that a finite percentage of the available ions (of the order of 1 - 5%) will be trapped into stable orbits and accelerated to the target. Thus, it has become clear that the well-known relativlstic limit to the energy obtainable from a cyclotron is not valid and the construction of machines to give many hundreds of millions of volts is practical. These principles have been tested on the 37" cyclotron at the Radiation Laboratory, under J. R. Richardson's direction, and a very satisfactory agree- ment between predicted and actual performance obtained. 4.0 Magnet — Structure The 184" magnet is similar in physical design to other cyclotron magnets. The pole diameter is 184" and by variation of the number of pole disks the use of different pole gaps is permitted. As at present planned the nominal gap used will be 24". The magnet is con- structed basically of 2" thick low carbon steel plate and is tied together by welding and bolts. Welds are purely structural and have no magnetic significance. About 4,000 tons of steel are used in the magnet structure. Operating conditions as planned call for a field strength at the center of about 15,000 oersteds. Variation of either or both the exciting power and gap will permit operation up to perhaps 16,000 - 17,000 oersteds. 4.1 Magnet-— Coils and Cooling The magnet is excited by twenty-two pancake shaped spiral coils containing approximate- ly 300 tons of copper and disposed in two coil tanks on either side of the gap. These tanks contain cooling oil which is circulated continuously through spray- cooled heat exchangers in the base of the cooling tower. The cooling oil also provides insulation between the coils and between coils and ground. 4.2 Magnet Excitation and Control As at present planned the magnet will be excited by two motorgenerator sets independent- ly regulated, one of 550 KVA and one of 360 KVA maximum capacity. These generators excite independently groups of coils symmetrically disposed in the two tanks and are subject to independent control and current stabilization. With the connections proposed, the maxi- mum power delivered to the magnet will approximate 750 KVA. This generator arrangement -4- . MDDC-983 is solely davised to use available equipment. It is necessary to maintain the magnet current constant at any arbitrarily selected value within about 0.01%. This is accomplished by means of electronic regulators such as is shown on schematic 2V 5844. A great variety of regulators are possible; this imit may be regarded as typical of tliis type of equipment. 4.3 Other Magnet Arrangements Basic magnet construction may use castings or forgings of appropriate quality as econom- ic considerations dictate. Magnet coils may be water or air cooled when considered desirable. 4.4 Magnet— Shims It is necessary for the operation of any cyclotron that focussing be provided to ensure that the ions move in stable paths from the ions source to the target and are not lost to the walls. This can be accomplished by providing a azimuthally symmetric magnetic field which decreases in a predetermined manner from the center to the periphery. For the 184" machine a total decrease of 4 to 5% is planned and over the major part the decrease will be linear. For the 37" machine as used for predicting the performance of the 184", an additional decrease was incorporated to simulate the increase in mass of the ions which will result on the large machine. The desired magnetic field shape is achieved by suitably machining the disks which form the pole faces of the magnet. These contours are determined by measurements made on scale models. Provision for empirical minor adjustments of the field on the 184" magnet are provided by spacing the pole face disc away from the core by 3/8"; iron shims can be placed in this gap as required to further correct the field. In addition, the peripheral and central parts of the shim contours are machined on removable sections which can be readily altered should occasion demand. Connections to the magnetic field might also be made by the use of a system of con- ductors carrying- electric currents. Electrical focussing might also be used to replace or supplement magnetic focussing — such focussing may be provided by the geometry of the accelerating electrodes or by estab- lishment of potential gradients within the vacuum chamber. 5. Vacuum Chamber — Structure The vacu\mi chamber for the 184" cyclotron is of welded steel construction and consists essentially of a flat square box with circular holes on top and bottom through which pass the magnet poles with suitable gasket vacuvim seals. Two of the sides of the box are closed by single large cover plates from one of which! the accelerating electrode or "dee" is[supportedand removed integrally with the plate. A number of large ports are provided on the other two -5- MDDC-983 sides for various purposes. Vacuum compression forces are carried by the magnet poles, by fixed or moveable col- umns at the large ports, and by external I beams across the corners of the top and bottom. Nonmagnetic materials (e.g.,brass or bronze, aluminium alloys, nonmagnetic steel) have commonly been used for such chambers with appropriate methods of fabrication. For small- er machines bronze castings have proved suitable despite difficulties with vacuum problems. 5.1 Vacuum Chamber — Vacuum Seals Vacuum seals for removeable and semi-removeable parts are in general made by means of rubber gaskets. The use of such gaskets was well established before the war and modi- fications made during the past history of the project have been well documented. In many of the Joints used at present it is desirable to have metal to metal contact between the metal- lic surfaces for reasons of alignment or electrical conductivity; this is accomplished by suitably relieving the gasket grooves so as to permit such contact when the gasket is suf- ficiently stressed to result in an adequate seal without permanent deformation of the gasket. In general, use is made of double seals with a test volume between them communicating to the outside to facilitate leak hunting. For certain applications modifications of our standard practice have been made to resolve particular problems. Thus,the gasket seal between the vacuum tank and pole disk is made by a heavy square rubber gasket compressed against both the disk and the tank by means of a pressure ring. In certain locations, use continues to be made of pipe thread joints (using a suitable seal- ing compound) and soft metal gaskets (such as aluminum or copper) under very high pres- sure. When relative motion of parts is desired, metal bellows are frequently employed and in some applications flexible metal hose. For rotating seals and sliding joints of long stroke, the use of rubber seals of the "Wilson" type are commonly used. Our present practice is to use "Chevron" seals when the standard sizes in which these are available are suitable. The application of these seals, originally designed for hydraulic service, to vacuum use was made during the war. Frequently the guard spaces between multiple seals are filled with diffusion pump oil for better sealing and to provide lubrication. For electrical connectors of low current and moderate voltage the use of mica insulated bushings is continued. Such bushings, similar to aircraft spark plugs, are available from the B and C Corporation and certain sizes are made to our specifications and drawings. For vacuum testing, use is made of helium leak detectors, either of the commercial type ^B^., that manufactured by the Westinghouse Elect. & Mfg. Co.) or the similar equip- ment developed at the laboratory. Pressure testing, Freon testing,and testing imder vacuum with gases (such as natural gas) which effect the sensitivity of a vacuum gauge are also employed. -6- MDDC-983 5.2 VacuuK) Chamber — Vacuum Pumps For evacuation of the main vacuum chamber and auxiliary systems (such as the vacuum condenser to be described later) use is made of oil diffusion pumps of the design developed during the war for use in the electromagnetic process. Thus, for the vacuum chamber of the 184" machine two thirty-two inch pumps with associated eight inch backing pumps are used, while for the vacuum condenser a twenty inch pump in connection with a similar back- ing pump is used. These pumps are in turn backed by a mechanical vacuum pump or pumps of suitable size. In general, mechanical pumps manufactured by Kinney, Beach-Russ, Cenco, and Welch are used in this laboratory as the service demands. The oil used in the diffusion pumps is commonly a petroleum distillate such as is manufactured by Litton Engineering Laboratories or the Distillation Products Co. Another useful pump fluid is "Octoil", and the use of silicone oils or fluorocarbons will probably become very common. Provision is usually made to incorporate a vacuum tight valve between the diffusion pumps and the vacuum system; such valves may be manually operated or remote control systems using electric motors, or air or hydraulic pressure, are occasionally used. Suit- able limit switches may be used to indicate centrally the position of all valves. 5.3 Vacuum Chamber — Lining Large electrical charging currents flow over the interior of the vacuum chamber as on the accelerating electrode structure. To reduce power losses and subsequent heating to a minimum, copper linings or copper plating is used of a thickness appropriate to the skin depth of the current. Water cooling as necessary is used on the lining and connections. "It is necessary to take suitable precautions to ensure that all joints be of such a nature to present low impedance at the frequencies used. 6.. Accelerating Electrode — Structure Either one or two accelerating electrodes, or "dees" are used in current cyclotrons. In general, a single dee is used for frequency modulated cyclotrons since the required energy increase per revolution of the ion does not require excessive voltages on a single electrode. Under certain conditions the use of more than two dees might be desirable although no such machine has as yet been constructed. The method of construction and support is dependent to a large extent upon the size of the unit and the type of radio-frequency system to be used. In the case of the 184'*" machine, the dee structure comprises a riveted and bolted dural frame structure which carries a covering of water cooled copper of thickness about 3/32". In view of certain experiments which indicate increased electrical breakdown from aluminum alloys, care is taken to completely protect the dural surfaces from direct exposure to volt- age gradients and possible sputtering action of the ion beams. The dee is approximately semicircular in shape and is supported from the circumference through a point opposite to the diametral edge. A slot is provided through one side of the support structure through which the deflected emergent beam can be directed. Since the water cooling tubes are at- tached to the interior surfaces to reduce danger of water leaks resulting from damage to -7- MDDC-983 the tubes from mechanical causes or electrical arcs, provision is made to separate the dei into two halves for servicing purposes with a minimimi of work. Bombardment of the in- ternal periphery of the dee by high speed ions results in considerable radioactxvity of these parts which may result in radiation exposure to servicing personnel. To reduce this, a channel is provided into which strips of graphite or other material with low induced radio- activity may be introduced and readily removed and discarded during servicing operations. Similarly the center of the diametral edge near the ion source is subject to considerable damage from ion bombardment and this section is made of heavy material, especially cooled and readily removable for replacement or repair. A general requirement upon a dee in addition to those indicated above is adequate rigidity and freedom from thermal distortion in operation. Additional considerations to be met by the design include provision for low resistance paths for the RF currents, smooth external surfaces to minimize sparking and the minimum possible electrical capacity consonant with adequate internal clearance for the ion beam. For smaller units, it is feasible to make dees entirely of sheet copper construction with a minimum of additional supporting members. Castings, stainless steel, or heavier copper sections may be used as well as dural for such members. As will be discussed later, it is sometimes desirable to incorporate a variable capacitor in the dee itself for producing the desired frequency modulation. Such a device is incorporated in the design for the Harvard unit and is tentatively planned for the machine imder design for Rochester. However, for the 184" machine the frequency modulation is produced by an external capacitor in a sepa- rate vacuum system. It is possible, and frequently desirable, to incorporate an ion deflection system within the dee. Depending on the type used, this may involve additional high voltage electrodes or magnetic devices. The Harvard and 60" designs are representative of this method. 6.1 Accelerating Electrode - Adjustment It is necessary that the dee be correctly positioned within the tank. In smaller imits it is usually possible conveniently to provide external continuous adjustments for this purpose which may be positioned during operation. However, experience has shown that such ad- justments, while desirable, are not essential, and it is adequate to provide means for ad- justing the dee accurately to a predetermined position. The latter practice is to be followed on the 184". Necessary adjustments require motion of the dee in all degrees of freedom. Since little ion source development for the low dee voltages experienced in a synchro- cyclotron has yet been carried through, imcertainties in die size of the source and its optimum, system of accelerating electrodes make it desirable to permit motion of the dee in and out of the tank (perpendicular to the diametral edge). Lateral motion of the dee may be necessary to secure additional orbit clearance for the deflected beam. The other adjustments are required to take care of fabrication inaccuracies, and inevitable flexures and other distortions. 6.2 Accelerating Electrode-Support The support system is intimately bound in with the radio frequency system adopted. A .general description wlll.be given here of the methods used on the several machines and the -8- MDDC-983 reasons for their adoption will be clear from the later discussion of the resonant systems. On the 60" constant frequency cyclotron, each dee is supported at the end of a copper covered steel tube supported from, and adjustable through, gimbals at the end remote from the dee. Since this tubular support with its associated copper lined steel tank forms a quarter wave resonant line at the frequency u^ed when loaded with the dee capacity, the gimbals can be rigidly connected to the steel tank without the use of insulators. Resonant frequency is adjusted by altering the position of a grounding spider which connects the dee stem to the tank lining between the gimbal and the dee. It will be noted that this construction does not permit placing any bias voltages upon the dee structure and avoids the use of any insulators in the main support system. When an electrostatic deflector system is used with the dee, the supporting member and current lead can be placed within the dee support tube, obviating the need for radio frequency by-pass capacitors of large current capacity. A somewhat similar system is under design for the Harvard cyclotron. In this case,two stems in parallel support the single dee and together form a quarter wave system. The mechanical frequency modulation xmit is mounted between the two support stems and pro- vides a variable impedance between the dee and ground. Again provision is made for an electrostatic deflector by the use of the same principle of complete radio frequency shield- ing described for the 60" case. While the design is not so far advanced, it seems probable that a similar system will be used for the Rochester cyclotron. The 37" synchro- cyclotron supports the dee through a single ceramic insulator from a single stem. In this case,the dee voltages are not large (circa 15 KV) and no difficulty is experienced with an adequately cooled and mounted insulator. The size of the unit is such that no especial mechanical problems arise from such a support. The insulator system permits bias voltages and this use has proved desirable. The 184" machine is designed to use an equivalent circuit to the 37" and the modifica- tions in the support structure result from the increased size and weight involved together with the increased electrical capacity of the dee. The dee is supported from four insulators so arranged as to be under purely compressive loading under operating conditions. (Mode- rate tensile loading on two insulators exists when the vacuum chamber is not evacuated.) These four insulators are located near the center of the dee support plate and lead tafour resonant lines which join the dee and variable capacitor and form a half wave resonant system. The large electrical capacity requires the use of multiple lines or a single line of such low impedance as to be difficult to achieve. Thus the mechanical and electrical considerations both lead to a multiple support structure. It is to be noted that these in- sulators are located quite close to the voltage node on the resonant system (the portion of this node^hifts of course as the frequency is varied) and as a result the insulators are only subject to moderate voltages at all times. Provision is made to equalize the mechani- cal stresses on the insulators. Clearly a wide variety of support structures are possible either using the voltage nodes of resonant systems or using insulators whether interior or exterior to the vacuum system. It is also possible to support the dee from ground through heavy springs which act as chokes to the high frequency voltage. -9- MDDC-983 7.0 Ion Source The ion source usually used in cyclotrons consists of a specially adapted low voltage gaseous arc at the center of the cyclotron. Such arcs have been frequently described (cf. the articles by Livingston cited earlier). Additional accelerator electrodes mounted on a dee are often used in increasing the electric field gradients in the neighborhood of the arc plasms. An ion source of this type is, when properly operated, an efficient source of doubly charged helium ions for alpha particles. For certain experiments pulsing the ion source in synchronism with other equipment is desirable. Some work has also been done on the development of special sources for production of multiple charged heavier ions such as beryllium and carbon. The use of high vacuum spark and especially heavy pulsed currents appear desirable in this connection. 8.0 Resonant System -General Requirements The radio frequency system for a synchro- cyclotron differs from that of a constant fre- quency cyclotron by the requirement of variable frequency. The extent of this variation is determined by the final energy to be obtained, (and corresponding mass increase of the ion) and the radial decrease of the magnetic field necessary for ion focussing. In addition it is desirable to provide additional variation (by perhaps 50%) since there is evidence that the optimum shape of the frequency-time curve can be more easily achieved if this be done. For 200 Mev deuterons and the 184" cyclotron, a swing of 25% is at present considered desirable. In view of the considerable amounts of radio frequency power required, considerable and worth-while gains in efficiency at these frequencies are possible by the use of resonant lines and cavities as circuit elements in place of lumped values. Considerable mechanical simplification also results from such use. Conventional cyclotrons have been driven both by self-excited oscillators and by power- amplifier systems. The former may be so arranged that the dee circuit itself comprises the main oscillator circuit or a more or less independent oscillator may be coupled to the dee circuit by transmission lines. In general, it has been this laboratory's experience that a self- excited system with the dee circuit as dominant resonant element is the most satis- factory. The incorporation of wide frequency variation introduces new elements into the situation. The use of a variable frequency power amplifier such as is used in f.m. broad- casting is probably feasible, but requires a load circuit of extremely low "Q" which makes power requirements difficult to achieve. It appears at present that the most practical pro- cedure is to use tbe dee circmt as the principal resonant circuit and to modify mechanically the natural frequency of this circuit. Under these conditions, frequency variations of 25% appear easy to obtain without excessive power requirements to reach the desired voltages. It appears probable that the modulation range can be substantially increased when necessary. In such a system, the dee voltage tends to vary with the frequency change; this is under con- siderable control by proper design of the oscillator circuits. It is comparatively easy to arrange matters so that the dee voltage increase as the frequency decreases and theoretical considerations indicate that this is desirable. It is of course possible, by variation of the oscillator plate voltage, for example, to produce any amplitude modulation desired. Howevei^ present plans provide only the rising voltage characteristic mentioned above. -10- . MDDC-983 The simplest oscillator arrangement is probably the grounded grid circuit. In-this ar- rangement, the oscillator grid is grounded with respect to radio frequency voltages, and the plate and filament circuits consist of line sections inductively coupled to the dee circuit resonant lines. It will be noted that this arrangement requires that the tube be quite close to the magnet — should the magnetic field be too strong to permit normal operation with reasonable magentic shielding, the use of transmission line coupling may be indicated. The former system has proved adequate for the 37" and 184", while the latter is planned for the Harvard cyclotron. It is usually desirable in such circuits to operate the tube with anode grounded (DC not RF)and filament at high negative potential; a variable capacitor may be used in the filament circuit to adjust for optimum phase relations between filament and plate circuits. To minimize radiation, complete shielding of the oscillator system is necessary. It should be noted that a wide variety of arrangements to secure the necessary dee volt- age are possible and different systems will be used according to special requirements. In particular, the acceleration of protons to high energies requires the combination of high frequencies, large accelerating electrodes,and large frequency variations. For a cyclotron of size comparable to the 184" these considerations may make it desirable to use rescmant cavities as circuit elements and to build the oscillator tube or tubes integrally into the dee structure. However, it appears that further study will allow the use of conventional tubes in this case as well. 8.1 Resonant System- Frequency Modulator As indicated above, present practice is to use a motor driven variable capacitor to pro- duce the frequency modulation. Other methods of achieving the desired end are of course possible in principle at least. Thus, the use of variable inductances, either air or iron cored, employment of capacities using dielectric materials either stationary or in motion, as weU as the standard f.m. broadcasting techniques, may be mentioned. Some experimental work has also been done on the use of ion plasmas as variable elements, controlled by varying the ion density, with considerable promise. However, for the present at least,the use of a rotating capacitor in vacuum apparently meets requirements adequately and simply. The vacuum condenser used on the 37" cyclotron has been completely described in re- port RL 36.6.3 which has been released for publication. The major differences in the unit under construction for the 184" result from the larger capacities involved and the radically greater power and voltage planned for the dee circuit. Thus,a number of rotor disks are used, mounted on a common shaft, clearances are increased, the coupling Capacity is a multiple disk unit, water cooling both of rotor and stator are employed, a brush system carries arbor charging currents aroimd the ball bearings, insulators supporting the stator are external and of larger size and cooled by forced air and the interior of the housing is copper plated to reduce heating. For the Harvard, and possibly the Rochester cyclotrons, it is planned to use a capacitor mounted directly on the rear of the dee. In principle, such a unit will-be similar to those described, an additional requirement being attention to eddy currents resulting from rotat- ing parts in the magnetic field. -11- MDDC-983 through appropriate shaping of the condenser blades or by moving one set of stator blades relative to the others. Variation of the position of the mean frequency is possible either by altering the length of the "half- wave" lines or by the insertion of additional capacity or inductance in the circuit for the 184" or 37" cases, or by altering the portion of grounding spiders on the "quarter wave" lines planned for Harvard and Rochester. 9. Ion Deflector Systems For many purposes, it is desirable to direct the high speed ion beam outside the cyclotron chamber. In a conventional cyclotron, this is done by means of an electrostatic deflector. In this system, a channel is provided in which a nearly radial electric field is established to counteract 10-30% of the radial magnetic force. The dividing strip at the entrance to this channel must be thin to prevent loss of a large fraction of the ions. The separation be- tween successive ion radii depends on the energy gain per turn and in addition that result- ing from any procession of the orbits. Due to the latter effect,it is found practical to use this system for a synchro- cyclotron, and it is used successfully on the 37" machine. Studies of a number of other possible methods are being or have been made. Three of these will be briefly discussed. The use of a pulsed electrostatic deflector instead of the constant potential type presents certain advantages resulting from the higher voltages possible and reducing the disturbing effects of electric fields on preceding orbits in certain arrangements. In particular, initial deflection parallel to the magnetic field with subsequent radial increase has been considered in some detail. It is possible to weaken the magnetic field locally by the use of a magnetic shield or by the use of currents in a system of conductors. Thus, if the circulating ion beam enters such a shield it may be led outside the main vacuum system. Since it appears difficult to make such a weakened field sufficiently local in character, it may be necessary to provide other means, such as for example, an equivalent diametrically opposed disturbance, to prevent dis- placement of the ion orbits. Such a system has been used successfully by D. Kerst of the University of Illinois for betatron beam removal. It is also feasible to pass the beam through a thin scattering foil (of tungsten, for ex- ample) and introduce the fraction of the beam scattered out into either an electrostatic de- flector or a magnetic deflector of the type mentioned above. The use of expander coils, such as are used in betatrons, to move the ion orbits suddenly outward may also be of value. It appears probable that a final deflector system will embody one or -more of the princi- ples described, and that it may be expected that about 10 % of the circulating ion beam can be deflected by systems of these types. 10. Target Arrangements Possible target arrangements are of two types, those intended for application to the in- ternal current (commonly called " probe targets") and these using the deflected beam. The latter can also be passed through a thin metal window into completely external systems such as cloud chambers, scattering chambers, etc. For both internal and external targets -12- MDDC-983 arrangements must be made for rapidly changing targets without undue exposure to radia- tion, for the greatest possible flexibility in target types, and for metering the ion current striking the target. Probe targets commonly consist of an interchangeable target assembly on the end of an adjustable probe which can be adjusted radially ot the desired position. Withdrawal through a vacuum lock permits interchange of targets without loss of vacuum in the main chamber. Wilson or Chevron seals are used for sealing the sliding shaft, and water and current leads are provided. Metering the ion current to such a probe target presents considerable dif- ficulty, and the use of proper electron traps to stop secondary electron emission and steps to minimize random plasma currents are necessary; measurement of the power delivered to the probe through temperature measurements on the cooling water is sometimes used. The deflected beam is commonly passed through a vacuimi lock into a target chamber where it is incident upon a removable plate, A number of interchangeable plates are pro- vided so that different targets may be rapidly placed in position or the beam passed through a thin window into external equipment. Means are usually provided for automatically clos- ing the vacuum lock upon failure of a window or other accident. For high current cyclotrons the cooling of the target presents considerable difficulty, and in many cases it is desirable to spread the beam over an extended target area for adequate cooling by suitably inclining the target to the ion beam. In general it may be said that each bombardment is a problem in itself and requires special technique dependent on the nature of the target material and the results desired. 11. Shielding Since a cyclotron is a source of considerable amounts of gamma and neutron radiations, it is necessary to shield the operators and other personnel from possible harmful effects. This is accomplished at the 60" building by the use of water tanks of about four feet thick- ness with the use of small amounts of lead in certain locations. Other commonly used shield- ing materials are concrete, either poured or in blocks for easier handling. For shielding against neutrons of very high energy, it is probable that preliminary shields of lead or some other element of high atomic number may prove the most effective, backed up by consider- able thicknesses of concrete, wood, iron, or other suitable material. It may be noted that a great deal of work on shielding has been carried out during the war in connection with pile construction. Since parts of the machine exposed to direct ion bombardment or to high intensities of gamma rays or neutrons become intensely radioactive, it is necessary to provide portable and temporary shielding, as needed, to carry out repairs on the equipment. Remote control handling of targets and probes is desirable for this reason and is commonly used. 12. Control Circuits As a moderately complicated piece of electrical equipment, a cyclotron requires an ex- tensive control system to properly monitor and actuate its parts, for protective purposes -13- MDDC-983 both for the equipment and personnel, and to control, measure and regulate its performance. Since many of the components require considerable amounts of power, much of this equip- ment is of an "engineering" nature, while many special electronic circuits are used for monitoring and control purposes. A brief review will be given of certain aspects of the control problem to indicate the scope of the equipment and certain of its functions. Vacuum control equipment involves the control of the mechanical pump motors, diffusion pump heaters (thermostatically controlled in some cases), and remote control of diffusion pump gate valves. In vacuum metering,use is made of thermo-couple gauges, ion gauges, PIG vacuum gauges, and helium leak detectors and mass spectrograph type gas analysers. Protective equipment includes devices to cut off power, close valves, etc. in case of a serious vacuum failure during operation or under standby conditions. Magnet control equipment includes the motor- generator sets and their regulation as de- scribed earlier. Cooling equipment includes circulating pumps for water and oil and suitable electrical interlocks to give alarm and protect equipment in case of general failure or stoppage of an individual circuit. Oscillator equipment includes the high voltage rectifier for the tube plate power and other associated equipment for filament heating, etc. It is necessary to regulate the plate voltage within wide limits and to provide power limiting devices (such as current limiting tubes or monocyclic networks) to protect the equipment in addition to conventional overload devices. Variable speed drive is used on the vacuum condenser (e.g., by "thymotrol" control) and electrical tachometers indicate the operating speed. The usual metering, controls, and protective devices of a high power radio-oscillator are also necessary. An electrostatic deflector requires a high voltage rectifier (100 -200 KV) and control equipment for varying its output voltage within wide limits. Protective interlocking is of course also required. Remote control positioning of electrodes is desirable. The ion source Involves equipment for electrolyzing heavy water to obtain deuterium, remote control shutoff and regulating valves, and remote control adjustment of the ion source position is desirable. Power requirements involve cathode heating and regulating equipment and arc power. Equipment for accurately pulsing the ion source at the proper moment in the frequency modulation cycle and of controllable pulse length is necessary for certain experiments. Such equipment in many respects is similar to well-known methods used in radar. Typical of special metering equipment are units to measure the time of flight of ions from ion source to target, the duration and shape of the current pulse and its location on the frequency-time curve. Instruments for monitoring the general radiation level and for determining safe condi- tions in specific operations are of course essential. UNIVERSITY OF FLORIDA 3 1262 08909 7389 1