Elementos que componen la instalación 1 Red de agua fría

For Phase 2, the screen camera pixel calibration was checked using established proce- dures; beam-line conditions, particularly the dispersion at the measurement position, were confirmed; and tomography experiments in both the horizontal and vertical planes were performed. An initial study of the effect of varying bunch charge on reconstructed phase-space was performed.

2.4.1 Screen Camera Calibration

If any camera lens corrections have been made to optimise screen focussing, maximum reconstruction accuracy will be maintained only if screen cameras in the Tomography section are recalibrated for Horizontal and Vertical magnification (in mm/pixel for the beam). This can be carried out using non-beam images, collected with the LED illuminators turned on so that the screen holder edges are visible, as in Fig. A.3. Software written in the ‘Mathematica’ code [42] is available for the calculation, which is based on the known dimensions of the screen. New calibration values are recorded in

an EXCEL sheet, stored in a standard location so that they are available to all users.

2.4.2 Horizontal and Vertical Tomography Experiments

Tomography in the Vertical plane is exactly analogous to the Horizontal case. When the capability to reconstruct both Vertical and Horizontal phase-space was added, the potential then existed for processing existing image datasets in both planes. However, quadrupole tomography scan magnet current settings for the projection angle range se- lected for horizontal scanning (using a horizontally-focussing quadrupole) give a range of angles for vertical reconstruction which are far from optimal, as illustrated in Fig. 2.19. The clear difference between horizontal and vertical phase-space reconstructions is a result of the very restricted, irregular angular range of projections in the vertical, the quadrupole tomography scan settings having been chosen for a uniform, almost complete coverage (153×1o) in the horizontal. On the other hand, the vertical range is non-uniform and only 23o for this example. Specific vertical tomography experi- ments were therefore planned, using a suitable defocussing (i.e. vertically-focussing) quadrupole, with current settings calculated appropriately for the magnet.

Vertical Data Processing The dimensions of the tomography transfer matrix set, which is prepared along with the experimental input data, need only be [2×2×n] where n = number of projections, if just horizontal processing is carried out, but must be [4×4×n] if the vertical elements R33n, R34n, R43n, R44n are included as well.

The first stages of processing, where complete (x, y) images are handled, are common to both horizontal and vertical cases. It is at the stage when the data is projected, either onto the ‘x’ axis (horizontal phase-space) or the ‘y’ (vertical phase-space), that processing differs, through to the generation of sinograms and their associated ‘position’ and ‘angle’ arrays.

Figure 2.19: Using data from a scan optimised for Horizontal phase-space reconstruc- tion, results are compared with the same data set processed for Vertical phase-space. Quality is poorer, as the Vertical projection angle range is much smaller & non-uniform.

2.4.3 Dispersion

In general the particles of a bunch do not all have exactly the same energy but are distributed about a mean. In a magnetic dipole field, the horizontal trajectory will vary with the energy; this is an example of ‘dispersion’. In the presence of dispersionηx, an

energy spreadσδ in a beam of emittancex with beta functionβx adds a contribution

to the horizontal RMS beam size σx, described by

σx =

q

βxx+ηx2σδ2 (2.15)

and this would be reflected in the reconstructed phase space. Dispersion has been designed to be≈0 in the tomography section, as there are two preceding dipoles whose effect is arranged to cancel out. The 2nd term in Eq. 2.15 may then be ignored; this condition has been checked by measurement.

Dispersion Measurement Based on a known dispersion at position AR1-1 on AL- ICE, its value has been estimated at screen EMI-3 in the tomography section, as shown

in Fig. 2.20. The procedure is to measure the deflection of the beam centroid when the energy is changed, in the region of known dispersion AR1-1; this gives a measure of the actual energy change. The energy difference dE can then be correlated with the linac settings used. With this data, together with the deflections dx observed for known energy changes, we can calculate the dispersion η at the position of interest EMI-3, using:-

η = E0×dx

dE (2.16)

whereE0 is the reference energy, in this case≈12 MeV.

Energy is adjusted by setting the accelerating gradients in the cavities LC1 and LC2 of the main linac, which is located upstream of the start of the EMMA injection line, indicated on Fig. 2.20. Using the valueη = 0.82 m (at AR1-1) established in previous ALICE calibrations, we obtainη =−0.06 m (at EMI-3).

This result was close enough to zero to indicate that no adjustment to the three quadrupole magnets indicated in Fig. 2.20 was necessary.

Figure 2.20: The dispersion η, known from previous measurements at AR1-1 in the ALICE ring, is used to calibrate the LINAC energy change, which in turn allows the dispersion at the position of interest (the EMI-3 screen) to be deduced.

2.4.4 Variable Bunch-Charge Experiments

After initial tomography work at an average charge per bunch of 40 pC, a series of experiments was planned specifically to investigate the effects of higher charges, as measured by the Faraday cup (see Section 1.4). Bunch charge is controlled by adjust- ing the attenuation of the photoinjector laser beam which initiates emission from the caesium-activated photocathode. The starting point would be the minimum charge which could be reliably imaged on the screen, given the fixed exposure of the installed ALICE cameras. At the other end of the scale, the maximum charge is that obtainable with zero laser attenuation, the governing factor being the quantum efficiency (QE) of the cathode. There is a gradual deterioration in QE with current drawn and intensity of use, and consequently the upper bunch-charge limit varies, typically between 80 pC when ‘fresh’ down to 55-60 pC near the end of its usable life. The achievable charge

therefore depends upon the point in the re-caesiation cycle when the tomography ex- periment is scheduled.

Results of Experiments

Preliminary measurements using the quadrupole tomography scan technique, with QUAD-08 capturing images on screen EMI-YAG-03, were made to reconstruct hori- zontal phase-space.

Figure 2.21: A series of quadrupole tomography scans, with QUAD-08 on screen EMI- 03, has been made at a number of bunch charges up to the maximum obtainable. Reconstructed horizontal phase-space shows a gradual evolution with increasing charge.

An evolution of the phase-space distribution with increasing bunch-charge from 40 pC to 70 pC, is evident in Fig. 2.21. It is clear that in all cases the general shape and orientation is very similar. However, it is difficult to make any conclusions purely by inspection of the raw plots, which show significant noise in the regions surrounding the central core of the distributions. Methods of analysis which extract useful quantitative information are described later, in Section 3.1.1.