The 400 GeV

c£«A//s/3-PU Hie European Organization for Nuclearbelow Research, ground and the accelerator construction has CERN, has built a 400 GeV ,thus caused known minimum disturbance to the environment. as the or TheSPS. diameter It provides of the ring is 2.2 km, the maximum the physicists of Europe with world-classwhich could facilities be accommodated on the available site. for research in . Although the construction programme initially Construction has been financed by elevenenvisaged CERN a peak energy of 300 GeV, it proved pos­ Member States — Austria, Belgium,. Denmark,sible, Federal by using accelerator magnets with a high peak Republic of Germany, France, Italy, Netherlands,field, to increase the energy to 400 GeV. The de­ Norway, Sweden, Switzerland, and the Unitedsign Kingdom. intensity is 1013 per pulse at a pulse Tlie total estimated cost, spread over repetitionan eight-year rate of aboutevery one 6 s. pulse construction programme which began inThe 1971,parameters is of energy and intensity are tiie key 1150 million Swiss francs (at 1970 costs).features Protons of an accelerator from the point of view were first accelerated to full energyof in thethe experimentalSPS in programme which it can sustain. 1976 but provision of the full range of machine facilities, particularly, in the experimentalProtons areas, from the CERN 28 GeV proton synchrotron, is scheduled for completion at the beginningPS, are injected of into the SPS at an energy of 10 GeV. 1979, After acceleration they can be ejected towards two experimental areas (West and North). Tire West ex­ TL'? accelerator is built undergroundperimental on an ex­ area is scheduled to receive particles tension of the previously existingearly CERN site in located1977, and the North experimental area on the Franco-Swiss border near Geneva. about The site 18 months later. The areas are being equipped area now comprises 109 hectares in Switzerlandwith a range and of hadron, , , and 451.5 hectares in France. Tire ring tunnelbeams, in which many of them of the highest energy, intensity, the machine is built is an average of aboutand quality40 m ever achieved.

History of the SPS project

Tire coming into operation about 20out years by agohigh ofenergy physicists and accelerator ex­ accelerators which could achieve particleperts energies from throughout Europe. This resulted in 1967 in the GeV range, opened up a new era in ourin a secondunder­ ECFA report which strongly recommended standing of the nature of matter. Asthe so building often of a 300 GeV proton synchrotron on the happens, this increased knowledge provokedlines aof series the CERN design study of 1964. of further questions, and it was obvious that many of them might be answered by access to stillBy thehigher end of 1968, letters of intent to sup­ energies. In the early 1960s, followingport the re­the project had been received from six Member markably successful operation of theStates, CERN 28 GeVand a Project Director was appointed by the Proton Synchrotron and its twin, theCERN 33 GeVCouncil. The Project Director, together with Alternating Gradient Synchrotron at Brookhavenexperts from USA, CERN and Laboratories in the Member it became clear that the techniques used inStates, tire set about revising the 1964 design to in­ construction of these machines could corporatebe extended toimprovements in accelerator technology much higher energies. which had subsequently arisen. Meanwhile, the CERN Convention was ammended to take account of the ex­ In 1963 a committee (the European Committeeistence forof two separate Laboratories (the revised Future Accelerators, ECFA) was set up toConvention study the actually coming into force in 1971). whole field of high-energy physics in EuropeHowever, and authorization to of the project by all the nake recommendations concerning futureMember experimental States was not forthcoming because of dif­ facilities. It presented a report to theficulties CERN in the selection of the site for the new Council in the same year, in which the majorLaboratory recom­ and because of the high costs involved mendation was the construction of "a inneu' the proton project as it stood at that time. accelerator of a very high energy". In this situation, the Project Director pre­ In 1964 a detailed design of a 300 GeV protonsented in 1970 an alternative project for construc­ synchrotron was produced by CERN. It wastion envisaged of the accelerator. Tire crucial change was that the accelerator would be constructedthe proposalin a new to build the new machine alongside the CERN Laboratory elsewhere in Europe, andexisting the Council CERN Laboratory. This circumvented the invited the CERN Member States to proposeproblems possible of site selection and, at the same time, sites. Twenty-two site offers were received,greatly and reduced the cost. The savings came from about half of them were the object of a usingcareful the PS as injector, from using the existing examination by CERN to see if they met theWest criteria experimental area which already had large- laid down in the design study. scale particle detectors installed and, particularly, from using the existing "infrastructure" of the CERN From the end of 1965 a further studyLaboratory was carried (administrative and technical services,

0 J vic-j j;' ike CSRli give, 'jiih the 3PS ajoeler-

ator picked out in white, against the background

of Geneva and the Alps. The Franco-Swiss frontier

which crosses the site is also indicated.

(Photo Swiss Air)

etc.). The now project also included several op­ The participation of the full community of tions concerning future development of the accelera­ European high energy physicists in designing the ex­ tor. For example, by beginning with a "missing perimental facilities for the accelerator was en­ magnet" lattice of conventional magi.ets, higher sured from the outset. ECFA organized two major energies could have been achieved later by the in­ study sessions in 1971 and 1972, which established troduction of superconducting magnets in the gaps the basis of the experimental hall layout, the left in the lattice. available particle beams, and features of the major detection systems. In 1972 an SPS Experiments The new proposal was worked out in detail in Committee was set up to decide on the in it ia l ex­ the course of 1970 and presented as a two-volume re­perimental programme. This early start on preparing port "A Design of the European 300 GeV Research the experimental programme was necessary to ensure Facilities". It received support from the scienti­ that the large and complex detection systems needed fic community and in political circles throughout for physics at high energies would be built in time Furope. for the start of operation of the machine.

On 19 February' 1971, ten of the CERN Member As the design and construction of machine com­ States approved the "Programme for the Construction ponents progressed at CERN, it was realized that the and Bringing into Operation of the 300 GeV Labora­ technology of superconducting magnets had not ad­ tory". (Denmark joined at the beginning o f 1 9 72 .) vanced far enough for them to be incorporated in The project cost was set at 11S0 million Swiss the SPS. On the other hand, it iras found possible francs (at 1970 costs) over a construction period to build the full ring with convention?1, magnets of eight years. A new Laboratory, known as CERN which, given the ring diameter of 2.2 km and a mag­ Laboratory II, with its own Director General was net design allowing peak fields of 1.8 tesla, would set up to build the accelerator. (The two CERN give the accelerator a peak energy capability of Laboratories were unified at the beginning of 1976.)400 GeV. The construction schedule was fixed so as The s it e , which straddles the frontier, was made to provide beams into the West experimental area in available by France and Switzerland. The supply of the sixth year of the construction programme and electrical power was ensured by France and the into the second (North) experimental area in the supply of cooling water by Switzerland. The formal eighth year. design description of the accelerator, entitled "The 300 GeV Programme", was published in January The SPS first accelerated protons to the design i9~ :. energy of 400 GeV in lt'76.

3 Générai features of the SPS design

Several of the major features of theas SPS"continuous de­ transfer" and consists of debunching sign c*re set by the decision to locate thethe machinePS beam and peeling it off during ten turns by alongside the existing CERN Laboratory.a system The land inof kicker magnets, an electrostatic deflec­ this area is not flat, the surface heighttor, varying and a septum magnet so that a ribbon of even from place to place by several tens ofintensity metres, andis ejected towards the SPS. The ejection/ the bed rock underneath is a ridge of molasseinjection (a is done at a momentum which can be varied mixture of sandstone and marl) whosebetween depth varies 10 and 14 GeV/c. This is high enough to over the site from a metre to 30 metresavoid below problems ground of quality of the injection field in level. Since the SPS had to be built in thethe bedrock SPS magnets and to limit the necessary frequency for stability reasons, it was not possibleswing in to the use SPS radio-frequency accelerating sys­ the cut and fill method for making thetem. tunnel It also to reduces the time for which the PS is house the machine. It was decided to bore occupiedthe tun­ filling the SPS. This keeps the time for nel using a full-face boring machine suitablewhich the forPS is occupied with injection into the SPS sand type rocks. to a minimum so that it can continue to supply pro­ tons for its own experimental programme at lower The molasse ridge has a width which permitsenergy a and for filling the Intersecting Storage Rings. ring diameter of 2.2 km. (The maximum possible dia­ meter is desirable since it is one of the parametersHie beam from the PS first travels along a which determines the peak energy of the sectionsynchrotron.) of tunnel, TT2, which is part of the beam The depth at which the tunnel lies below transferthe surface system, towards the Ring 1 of the Inter­ on the circumference of this ring variessecting between Storage a Rings. Two hundred metres along minimum of 18 m and a maximum of 64 m. TT2 With a switching the magnet is powered to direct the minimum depth there is still more thanbeam adequate along a new tunnel, TT10, towards the SPS (a shielding by the overlying rock and moraindistance to of ensure 800 m). A septum magnet and kicker mag­ that any radiation produced in the tunnelnet isbend reduced the protons onto their orbits in the ring. to a level on the surface which is wellThe below ten-turn the ejection from the PS gives a ribbon of international tolerances for the generalprotons public. filling 10/11 of the SPS circumference. The magnetic field in the kicker magnet can thus The precise location of the ring wasdrop also to dic­ zero before the first injected protons pass tated by the fact that the proton synchrotronit on completing (PS) their initial turn so that their provides the particles for injection andtrajectories that the will not be disturbed by the injection accelerated particles are sent to the alreadyfield. ex­ isting West experimental area. This fixes the posi­ tions of two long straight sections, Theone magnet for injec­ system into which the beam is in­ tion (LSS1) and one for ejection to the Westjected area has two purposes — it has to curve the tra­ (LSS6). The ring is divided into six arcs andjectories six of the protons, so that they complete a long straight sections, the other straightno-t ~r.-cle sections around the ring, and it has to maintain being taken up for ejection to the Norththe experi­.cun well focused, so that the protons do not mental area (LSS2), for the radio-frequencyhit theac­ wall of the vacuum vessel in which they celerating system (LSS3), for the beam dumptravel. (LSS4), While being accelerated to 400 GeV, the and one left free (LSS5). protons orbit the ring about 150 000 times covering a distance of over a million kilometres. To ensure Six access shafts, located at the longthat straight they follow the desired trajectories throughout sections, are sunk to the ring, and sixthis sen/ice journey, it is necessary that the magnetic buildings are constructed on the surfacefields to behouse accurate to better than a few parts in the power supplies and other equipmentten feedingthousand the over the entire region in which the machine components installed in the ringprotons tunnel travel. un­ derground. There is also a limited zone for office blocks and an assembly hall, and too largeThe experi­ magnet system is of the "separated function" mental halls in the North area. The disturbancetype in towhich the tasks of bending and of focusing the existing environment has been keptthe to beam a minimum, are carried out ii\ separate units — dipole and farming and forestry continues onmagnets most andof quadrupole the magnets, respectively. This area of the SPS site as it had done previously.allows a higher peak energy than a "combined func­ tion" type for a given radius. There are eight di­ Protons are injected into the SPS frompoles the andPS two quadrupoles in each normal "period" of with a beam intensity which will enable thethe SPSmagnet to lattice. The pattern of this period is reach its design intensity of 1013 protonsi-epeated per 108 pulse.times around the ring. The quadrupoles In order to spread the protons aroundare thearranged SPS cir­ so that their fields are successively cumference, which is eleven times longerfocusing than that and then defocusing giving a net focusing of the PS (the PS diameter being 200 m), a noveleffect in both horizontal and vertical planes. method of ejection from the PS is used. Their It is peak known field gradient is 21 T/m. The bending

4 Plan of the SPS showing the link with, the P S w h i c h

serves as injector, and the two experimental areas. The six service buildings and access shafts which L/NAC are located at the six long straight sections of

the machine, are indicated together with a seventh, located over the beam line to the West Experimenio.l

A r e a .

DllU An ara o f t h e SPS ring s h o w i n g bending magnets a n d

fccrusing qnadîrupolss ir place. (SERS 10.11.74)

magnets are of the H-type in a compact candesign be ejected which to feed the experiments), and 1.2 s keeps cost and power requirements tofor a minimum. the magnet field to fall again. The power dis­ They are 6.26 m long, and each normal periodsipated has in the magnets is absorbed by a water- four magnets with an aperture of 145 coolingmm x 35 mm system. and Water is pumped from the Lake of four magnets with an aperture 120 ran Geneva* 48 mm. (Lac TheseLéman) and stored in two reservoirs, dipoles are positioned so as to match withthe undula­ a total capacity of 10 000 m3, on the SPS site. tions in the bean profile determined byWhen the thefocusing machine is in operation the necessary flow magnets. The dipoles are pulsed to giveof a coolingpeak wateri/s. is 1000 field of 1.8 T (corresponding to 400 GeV) with a peak current of 4.9 kA. The vacuum system is maintained at a pressure of better 10”7than Torr by about 650 sputter In addition to 744 dipoles and 216 quadrupoles,pumps and 80 turbomolecular pumps. The vacuum tube, there are correction magnets — dipoles,of sextupoles, low permeability stainless steel, changes in pro­ and octupoles to compensate for anyfile remaining as it threads de­ through the magnets and othe. fects in the magnet lattice and to allowmachine refined components. control of the orbiting beam. Acceleration of the protons is assured in The magnets are powered using a "staticstraight power section LSS3 by two radio-frequency cavi­ supply" rather than a motor-altematorties, set 20 m common long, each with a waveguide structure of in older . Thisbecause is thepossible 56 drift tubes. This form of RF structure is novel power generating capacity of the publicin electricitythe acceleration system of a synchrotron ring network is much greater than the peak andpulse is the power most economic way of transferring power required by the SPS, and the load on the generatingto the beam. Since the proton velocity varies very stations can be uniformly distributed.little The static (0.4$) during the acceleration cycle, it is supply allows the pulses of powerpossible to flow to between use an almost constant frequency of the network and the SPS and back again without200 MHz forsig­ the RF. This frequency divides the or­ nificant fluctuation of the networkbiting voltage. beam into It is 4620 bunches. The RF system ac­ necessary to connect at an electricallycelerates strong the protons by about 2.5 MeV per turn, point of the network and this is donecorresponding at Génissiat to an acceleration rate of over in France. A 380 kV line brings the power100 from GeV/s. Génissiat to CERN. The accelerated beam can be extracted towards The peak power for the bending magnetsthe West and area and towards the North area using ex­ quadrupolcj is 135 MW and there is a mean tractionpower systems located in straight sections LSS6 consumption of about 34 MW. This allowsand LSS2,an SPSrespectively. Three methods of extraction cycle time of about 6 s, including 0.2 scan for be operatedinjec­ to meet the different needs of the tion, 3.7 s for magnet field rise (accelerationexperimental programme. Fast extraction involves time), 0.7 s for a "flat-top" (during whichbending protons the protons out during a few microseconds The SPS Main Control Room, The accelerator nay be

operated from any of the three larger consoles /ehe

smaller one being used for safety controls.

(CEBU 137.10.75)

(all, or a fraction, of the bunches extractedFrom during the extraction point in LSS2 for the North one turn) and will be used particularlyarea, to proton provide beams can be taken almost 600 m to the a neutrino beam and a RF separated beam tosurface the BEEC and split between three targets to yield . Medium extraction involveshadron bending beams in North experimental hall 1, and the protons out during several milliseconds,hadron and muonwhich beams into North experimental hall 2. is the maximum duration of beam that BEBCThe cansecondary cope beams can be produced by protons with with while still being acceptable to someincident electronic energies up to 400 GeV. experiments. Slow extraction involves bending the protons out during times up to a secondMonitoring or more, and control of the multitude of providing beams best adapted to the needsmachine of elec­ components and of the proton beams themselves tronic experiments. Medium and slow extractionsare carried out in­ by a computer control system which volve exciting a beam resonance which progressivelyhas advanced the techniques of accelerator control increases the radial oscillations of thein protonsseveral ways. so The system has 24 small computers — that they enter the units of the extraction13 distributed systems around the ring, 8 at the Main Control which deflect them out of the ring. Room, and 3 in the experimental areas. They are all capable of independent operation and in some cases Each extraction system comprises fourare "dedicated" electro­ to the monitoring and control of static septa 3 m long with septa made ofspecific molybdenum components (for example, one computer looks wires 0.1 mm thick spaced l.S mm apart. after Outside all aspects of the radio-frequency system). the wire septa are electrodes at a potentialThe computers of communicate with one another and with -300 kV. The field deflects the protons threesufficiently consoles in the Main Control Room by means of so that they enter the aperture, of septuma high-speed magnets, message transfer system. of which there are four, 3 m long with septa 4 urn thick, followed by another five, 3 m longMuch with attention has been given to the control septa mm.16 thick. philosophy so that data can be called up easily and presented in a way which is easily assimilated by From the extraction point in LSS6 forthe machinethe West operators. The operator views a number area, proton beams can be taken up to theof Westcolour ex­ TV screens which present the status of perimental hall and split between threecomponents targets to of the machine which can be selected with yield secondary beams. The dimensions ofa "touch the hall screen". The screen has 16 transparent can accommodate secondary beams producedbuttons by protons backed by a TV upon which different labels with energies up to 200 GeV. Alternatively,can be protonswritten by a computer. Almost all the hun­ with incident energies up to 400 GeV fromdreds the ofsame knobs and switches of a conventional con­ extraction point can be used on undergroundtrol room targets are channelled into a single knob whose to yield either secondary beams of a typefunction and energy is also controlled by the touch screen. A which can be selected by RF separators, orspecially neutrino developed computer language is used which beams which are aimed at the BEBC and Gargamelleallows changes to any computer program to be made bubble chambers and counter experiments.rapidly and easily. Published by the Publications Section, Document No.CERN/SIS-PU 76-02

European Organization for Nuclear Research

1211 Geneva 23, Switzerland May 1976