Efficiency of created new accelerating complex VEPP-2000 is appreciably defined by reliability of an injection system including linear accelerator ILU (E=2.1 MeV), pulse synchrotron B3M (up to E=250 MeV), conversion system and booster ring BEP (maximal energy of electrons/positrons is up to 900 MeV).
Several steps of improvement to existing particle production system are proposed to gain the bunch population of 1011 particles. Injection of beams into the storage ring is planned to be done in the horizontal plane into the long straight opposite to the RFcavity at full working energy (Figure 1).
Injection system assumes single-turn injection with pre-kick of a stored beam. Two traveling wave kickers are arranged into vacuum chamber inside two bend magnets on both sides of injection region.
Passage of the injecting beam through nonlinear kicker field was modeled with combination of multiples, then injecting beam was checked on stability in VEPP-2000 structure with tracking procedure.The result of such procedure is Acceptance (Figure 2) of the collider in the injection point, which appeared to be 5.5 cm·mrad.
Due to lack of space injection region is separated on two magnets: low field (2 T) septum and high field (3 T) additional magnet. Both magnets are designed on coaxial-like scheme with laminated yoke. Figures 5 and 6 shows magnetic fields and gradients of both magnets, their main parameters are listed in Table 1.
|Radius||100 cm||151 cm|
|Field||3 T||2 T|
|Current||54 kA||38 kA|
|Inductance||0.14 µHn||0.12 µHn|
In the first half of 2006 successful start of injection part took place. The main results of that work are listed below.
Deep modernization of control electronics and software of VEPP-2000 complex have been done. Power electronics of ILU's RF generator has changed as well. This allowed one to increase the reliability of work. Design parameters of ILU have been reached: electrons were accelerated to the enrgy 2.1 MeV, pulse current (40 ns) was up to 4 A. Figure 7 shows the oscillogram from Rogowski coil installed in ILU-B3M channel.
To deliver the electron beam three diagnostic systems in the channel ILU – B3M are used: Rogowski coil, inserted in/out luminophor probe in the final rotation magnet IM2 and the system of three secondary emission grid sensors. These allowed us to increase the efficiensy of beam delivety up to 100 %.
Measurements of beam profile with different focussing lengthes of the first pulse quadrupole IL1 allowed us to estimate the beam emittance (about 1.5·10-3 cm·rad) and the initial angular divergence of the beam ejected from ILU. We have decreased this divergence with the aim of solenoidal quadrupole IL1 and consisted Twiss parameters of the beam with the channel acceptance.
The change of horizontal dispersion sign before injecting in synchrotron B3M is done by long matching solenoid. This solenoid realises "minus" unity matrix in both directions, hense allowing one to change dispersion sing after pass through the solenoid. Matching of dispersion with the value optimal for injection have done with the help of secondary emission sensors.
Tuning up the injection includes: matching of transverse phase spaces of the beam with synchrotron acceptance; choice of amplitude and pulse time of kicker; corrections of orbit dipole distortions; choice of optimal values for betatron tunes and much more. It was needed to match acceleration gain and gain of the guid field increasing. That was done by using additional inductance in power line of booster iron. All these tasks have been successfuly solved and the beam current of 2 A in the first step of acceleration have been obtained.
In the Figure 8 one can see typical oscillogram of betatron acceleration in B3M. Time duration of betatron acceleration cycle is about 160–200 µs and it is defined by duration of "flat top" in "step" pulse.
One of the most complicated tasks was optimization of transition from betatron to synchrotron acceleration. Here we have some losses of particles that are out of RF separatrix. But we need to minimize these losses by tuning RF system. Specially developed for control the RF amplitude microprocessor unit allowed us to change the RF pulse shape in a flexible way: RF voltage can be changed from 2 kV to 25 kV in the middle of cycle. In the end of cycle the deep modulation of RF frequency is done allowing one to direct the beam in the ejecting kicker gap.
During tuning up the ejectin from B3M we have realized the need of dynamical corrrection of radial orbit just before ejecting from the synchrotron. Two special orthogonal pulse correctors were created: RV1 and RV2, that produce first harmonic of guiding field. Each corrector makes positive and negaive distortion in two opposite B3M quadrants.
Special system of damping the coherent synchrotron oscillations have been created. This allowed us flexibly select the phase delays in 16 time intervals of acceleration cycle increasing the stability of synchrotron acceleration as well as the current and position of ejected beam.
Magnetic system of BEP has 12-fold symmetry, so we can change the working point in a wide ranges. Choosen working tunes (Qx=3.43, Qz=2.7) are opimal for minimal shift of working point due to increasing the current in bending magnets and quadrupoles of BEP (connected in series and feeded with current 10 kA). Experiments in booster regime showed that with initial tunes Qx=3.43, Qz=3.2 (also good for injection) the working point is shifted down to the integer resonance Qz= 3.0 during the increase of current in coils. This is because of advancing saturation of D-quads family as compared with F-quads and dipole magnets. Hence we decided to change focusing structure of booster to minimize currents in correcting coils of quadrupoles at high energies.
Choosing Qx near to half-integer resonance simplify the organization so-called "pre-kick", that is needed for increasing the amplitude of kick. Tha latter is used for placing the positron bunch to the closed orbit. In using scheme of multiple storage of positrons pre-kick and kick are applied to the stored circulating beam sequentially in one turn, but injected bunch is kicked only once. When working near to the half-integer resonance pre-kick and kick cancel each other.
In the Figure 10 the oscillogram of injection into BEP is shown.
Conversion system of VEPP-2000 is based on pulse lithium lenses which focussing electron beam to the tungsten converter and collect positrons from converter. Electrons energy is 250 MeV, positrons – 125 MeV. The thickness of converter is 3 mm, this is about 0.8 radiation length. Angular divergence of positron beam is about 0.1 rad. The lithium lenses are of 1.5 cm in length, the diameter of lithium rod is 5 mm. The current of 100 kA is going through lithium rods, that makes linearly increased with radius magnetic field (up to 100 kGs at the rods surface). Focussing length of these lenses is about 1.5 cm, that allowed one to concentrate electron beam to the spot of 0.1 mm in diameter.
Optimization of positron injection in BEP includes precision alignment of short focussing lithium lenses, which are assembled on special platform with electric motor drive to move lenses in all directions to each others. Then by measured with photomultiplier single injection of positrons optimization of triplet's gradients and correction of injected beam trajectory was done.
Energy spread in collected positron beam is ±3%. Such big value required to use pulse modulation of RF voltage: main level (25 kV) is changed during short time (0.2 s) up to 60 kV. This was done to increase energy separatrix of RF.
Integral coefficient of conversion 3·10-4 is reached. This allowed us to store about 100 µA of positrons after one injection, that is enough to get designed luminosity of VEPP-2000.
In spring of 2006 the work on deliver electrons and positrons from BEP to VEPP-2000 has started. Tuning of channels is going successfuly, beam position monitors steadily see the beam. The ratio signal/background at ejected current of 30 mA is about 1000, this allows us to tune channels at currents of 1 mA.
To measure beam position along injection channels two types of beam position monitors are used: secondary emission sensor (SES) and image current monitor (ICM) (Figure 14). Figure 15 and 16 shows data from SES and ICM respectively.
Each vacuum chamber contains (in the middle cross section) a water cooled triangle mirror, which reflects the visible part of the synchrotron radiation from both beams. This light goes outside through a glass window to the optical diagnostic system: beam current (PMT) and beam position and dimensions measurements. CCD-cameras are used as beam position and size recorders in 16 points around the ring. In addition to optical BPMs there are 4 pick-ups in the technical straight section and one current transformer as an absolute current monitor.
In the present state all solenoids are produced and assembled in blocks together with iron yoks, vessels for liquid helium and nitrogen and outer shielding. All solenoids have been tested abd showed the desired field parameters. Two blocks of solenoids are alredy installed in the SND region and are ready for final testing.
Up to now a final assembling with vacuum vessel and LHe tank was done. Tests of solenoids installed at the storage ring showed a liquid helium consumption twice higher than a calculated one.
The internal tube of the helium vessel (dia. 50 mm) is a part of the collider vacuum chamber and thus provides for cryogenic pumping in the interaction straight sections. The synchrotron radiation from the bending magnets is intercepted by a nitrogen vapor-cooled liner placed inside.
High vacuum pumping of the experimental straight sections is performed by the internal tube of the LHe vessel housing the SC solenoids. In the rest regions, combined ion-pumping and getter pumping are used to cope with gas desorption from the vacuum pipe irradiated by the synchrotron radiation. Bakeable stainless steal vacuum chamber is equipped with water-cooled radiation cooper absorbers and have to provide vacuum 10-6 Pa at the beam current 2×150 mA.
Before commissioning of VEPP-2000 itself we had to restore the injection part of the accelerator complex. This work started in the early 2006 and was developed step by step following a readiness of corresponding control and supply systems along a chain of transfer lines and accelerators: 3 MeV linac ILU, 250 MeV synchrotron B-3M, buster storage ring BEP. This process reached the VEPP-2000 border near to the end of year.
In the storage ring VEPP-2000 solenoids are the basic focussing elements formativing the demanded aspect of beam envelopes in the observational gaps. Optics cruising regime β=10 cm because of affinity of an operating point to the whole resonances Qx=4.1, Qz=2.1 is characterised by high sensitivity to a direction of solenoids' optical axes. It is required to expose a rule of symmetry axes of a field with accuracy about 0.2 mm. Thus the amplitude of distortion of the closed orbit in working alternative of optical structure will make about 3.4 mm (the coefficient of amplification on orbit distortions is equal approximately k=17, diameter of a vacuum chamber is equal D=40 mm). Editing of orbit's distortions which are brought in by biases of solenoids becomes complicated unusual for functionals an orbit twisting. Development of such nonstandard optics is represented a challenge.
For initiating injection of bunches in ring VEPP-2000 it was offered to use auxiliary focalizing structure with completely switched out solenoids. Such approach allows to facilitate feeding into in a system of the basic systems of the accelerator (optical CCD matrixes, pickups-supervision, measurings of an oscilation frequency, RF systems etc.), and also to inspect working capacity of all injection devices (septum magnets, kickers). On the other hand the intermediate starting optics allows to do trial closings of all solenoids separately, in particular, to be convinced of correctness of their agency on twisting the distorted orbit, and also to yield rather coarse prestress mechanical alignment, using a bundle as the sensing transducer feeling a rule of an solenoid's axis. After editing of orbit distortions by available dipole zero adjusters, and if necessary by mechanical resliders of solenoidal modules, the gradient junction to the nominal optical structure with an operating point near to the whole resonance and small β-functions at centre of the observational gaps is carried out.
Such approach to initiating adjustment of the storage ring in practice has appeared very convenient expedient of reception of a circulating current. By measuring of several hundreds of bunch's first turns in a regime with RF cavity being switched out we measured circulating frequency in a store magnetic field f=12292 kHz which with accuracy better than 1 kHz has coincided with the design value. In this connection it was not required additional force for capturing of particles at turning on RF voltage. At the first powering on the RF the beam is captureed lost-free of intensity in a synchrotron regime of acceleration.
This "soft" optics (Qx=1.2, Qz=2.4) strongly enough differs from structure with a round beam. But thus the structure part in an injection gap is close to the design. The first circulating electron beam has been captureed on energy 140 MeV, and soon and on energy 508 MeV. On energy 508 MeV which on that time was restricted to the power supply of the basic magnets, control systems, beam diagnostics and corrections have been mustered and colibrated.
After obtaining stable capture of a beam there was possible to measure the parametres of VEPP-2000 ring optical structure.
Measuring the chromatism of betatron tunes became one of routine points of studying the ring optics. Dependence of betatron tunes on frequency of the treatment has been taken at switched out regular sextupoles (Figure 20) which shows the considerable discrepancy with the design value of natural chromatism. Probablly, this discrepancy speaks that magnetic permeability of a massive vacuum chamber in dipole magnets differs from unity and, as show precision determinations, makes μ=1.005. The vacuum chamber configuration creates feeble nonlinear components of magnetic field, including sextupole one. Because of major extent of magnets the contribution of this not considered sextupole in chromatism of betatron tunes appears appreciable. Nevertheless, this nonlinearity, owing to it's smallness and distributions on a ring, should not influence a size of the dynamic aperture, and value of chromatism is easily cancelled by regular sextupoles both in a regime of adjustment of optics, and in an operating mode with round beams.
Orbit measuring in collider VEPP-2000 is carried out by means of two systems: pickups-stations (on one on the quadrant) and CCD matrixes (16 units, by 8 on the electron/positron direction). In Figure 17 the effect of on-line machining of the images gained from CCD in the form of coordinates and sizes of a bunch along collider structure is presented. The software created for system of CCD, allows to correct easily an orbit by means of prestressly taken response matrixes. From the figure one can see, that the mean deviation along an orbit makes 0.2 mm, and the size will well be compounded with design data.
When the capture effectiveness has attained 70-80 % the bakeout of a vacuum chamber by a synchrotron radiation of an electron beam in both directions has been yielded. After several days of agings the beam current has attained 150 ěŔ, and life time 1000 s. Under these requirements the life time of a small current (the order 1 ěŔ) exceeds 10 hours.
VEPP-2000 operation without solenoids took about half a year. It was caused mainly by a low capacity of the liquid helium production at BINP, which was not enough to keep simultaneous experiments at VEPP-4M collider and solenoids feeding at VEPP-2000. Only at the end of May 2007 the cryogenic system of VEPP-2000 was put into operation.
First of all we had to prove an alignment of the cooled solenoids. It was done by the CO deviation measurements as a response to the orbit steering coils, first performed in the same “weak focusing” regime. Each section of all 4 solenoids was tested with magnetic field level up to 4 T. So, coordinates of each i-th solenoid section center (xi, zi, x′i, z′i) have been obtained from the Orbit Response Matrix analysis (ORM), and necessary mechanical shifts of the solenoids have been done. After this preliminary alignment the simplest round beam regime (+ − + −) was applied with 1T field in the anti-solenoid and 10 T in the nearest to IP section of the main solenoid. The round beam machine lattice for is shown in Fig. 21.
The electron beam was successfully injected just after the solenoids “on”, with fractional tunes near a half-integer, Δν1≈Δν2≈0.5. Later on, few steps of the CO and lattice functions corrections have been done aiming to bring the tunes near to integer. At that, the SVD method was routinely used to minimize a sum of currents in dipole steering coils and deviations in focusing strength of quadrupoles and solenoids from original symmetry. Finally, we get a regime with Δν1≈Δν2≈0.1÷0.15 and moderate CO deviations (Δx≈Δz≤±1.5 mm) from the axes of quadrupole magnets (see Fig. 22).
At that time one system of the storage ring was not complete. Instead of four kicker generators, only two of them were in operation. So, we could inject the beams, but without storage. Also the positron source worked far from its full strength. It was able to deliver the positron current of 3-4 mA only in the VEPP-2000. In this situation we came to a decision to make a round colliding beam test in the “weak-strong” option rather than curing some visible machine imperfections. The simulation of the “weak-strong” option predicts a weak dependence of the IP beam size on the opposite beam strength ξ (see Fig. 23).
Experimentally we measured horizontal and vertical beam dimensions of the positrons in other positions. Figure 24 presents the rms beam sizes at three points versus the electron beam current. In point 3 located in dipole nearest to IP, there is a minimum of βx seen in Fig. 21.
According to the simulation, the horizontal rms size behavior at this point (dark blue) is similar to the IP. The vertical size (light blue) grows as a result of the counter beam focusing, that βz increases and the radiation emittance.
From the task of automation's point of view VEPP-2000 complex represents more than 500 management channels and more than 1000 channels of the control. Sharing of such quantity of channels superimposes hard demands of a control system of the complex. The control system of accelerating complex VEPP-2000 is based on several PC platforms under control of operating system Linux, connected in the common local area network. The circuit of VEPP-2000 complex automation is presented in the Figure 25.
At sampling of the protocol for hardware dialogue with controlling computers it is necessary to be guided by standards, abundance, a support, capacity, and other important criteria. For automation of accelerating complex VEPP-2000 two main protocols have been selected: widely known and well proved in tasks of automation of scientific researches standard CAMAC and rather new protocol of industrial link CAN-bus.
Standard CAMAC on VEPP-2000 complex is used in those places where the big capacity is required, for example in channels of observation for a bunch, and also in subsystems where replacement by newer standards has been recognised by inexpedient. CAN-bus protocol is base in system of automation ÂÝĎĎ-2000. It is very convenient for spatially separated control systems and allows to reduce quantity of wire communications considerably.
Principles of construction the software on VEPP-2000 complex are grounded on the hardware architecture. Specialised servers inspect one or several buses CAN-bus or CAMAC and allow client applications to have access to management channels and the control, to appropriate these buses. Distinctive feature of such approach consists that some client applications can simultaneously fulfil handle and carry out monitoring and the control of hardware operation. Applications can exchange system-wide events and commands (for example, mode change, or change of working energy etc.) through a special server of messages. The second feature of this architecture of the software - hiding from the end user (operator) the details of operation with each concrete block of hardware and details of implementation of each specialised server. All information on a hardware configuration is stored in databases specially developed for these tasks, own for each subsystem. It is possible to illustrate all aforesaid with the Figure 26.