Beam feasibility study of a collimator with in-jaw Beam Position Monitors

At present, the beam-based alignment of the LHC collimators is performed by touching the beam halo with both jaws of each collimator. This method requires dedicated ﬁlls at low intensities that are done infrequently and makes this procedure time consuming. This limits the operational ﬂexibility, in partic-ular in the case of changes of optics and orbit conﬁguration in the experimental regions. The performance of the LHC collimation system relies on the machine reproducibility and regular loss maps to validate the settings of the collimator jaws. To overcome these limitations and to allow a continuous monitoring of the beam position at the collimators, a design with jaw-integrated Beam Position Monitors (BPMs) was proposed and successfully tested with a prototype (mock-up) collimator in the CERN SPS. Extensive beam experiments allowed to determine the achievable accuracy of the jaw alignment for single and multi-turn operation. In this paper, the results of these experiments are discussed. The non-linear response of the BPMs is compared to the predictions from electromagnetic simulations. Finally, the measured alignment accuracy is compared to the one achieved with the present collimators in the LHC.


INTRODUCTION
To intercept unavoidable losses of particles from the beam halo that would otherwise risk to hit the superconducting magnets, the Large Hadron Collider (LHC) has a powerful collimation system with 44 movable collimators per beam [1, 2,3]. Most collimators consist of two jaws, which can be moved 5 independently, with the beam passing through the center of the jaws. Each jaw is called 'left' or 'right' depending on its position with respect to the beam when viewed from the upstream side of the collimator. For optimal performance, the jaws have to be centered around the local orbit. This has so far been done using a beam-based alignment procedure for each collimator [4], where each jaw is 10 moved separately towards the beam until it starts intercepting the halo particles. This is verified by monitoring the signal of a nearby downstream beam loss monitor (BLM), which registers the secondary shower particles created by impacts on the collimator. For machine protection reasons, the alignment procedure requires dedicated fills at low intensities that are done infrequently because the 15 procedure is time consuming [5]. The introduction of a semi-automatic set-up procedure and constant improvements in the algorithms allowed to significantly reduce the set-up time in 2011 and 2012 compared to the first manual set-up in 2010 [6,7]. When all collimators have been centered around the beam, the cleaning performance is verified by provoked losses to create a so-called 'beam 20 loss map'. In subsequent high-intensity fills, the collimators are driven back to the previously found positions, relying on the machine reproducibility. This implies strict requirements on the long-term orbit stability, as the time-consuming setups cannot be performed too frequently. The excellent performance of the LHC collimation system during run 1 has recently been discussed in [8]. 25 To overcome these limitations, a new collimator design with in-jaw beam position monitors (BPMs) was proposed. Four BPM pickups are installed at the extremities of each jaw to provide a measurement of the beam orbit at the upstream and downstream sides of the collimator. Beam tests were successfully carried out with a mock-up collimator in the CERN Super Proton Synchrotron 30 (SPS) [9,10]. Figure 1 shows a photograph of the prototype collimator. The moveable jaws are centred around the beam path (red arrow) and enclosed by a 1.2 m long tank. A sketch of the mock-up jaw with the BPM pick-up buttons in the beginning (upstream) and end (downstream) of the jaw is depicted in Fig. 2. A BPM-based alignment, where it is not necessary to touch the beam with the collimator jaws, would allow a fast and non-destructive beam-based colli-  mator set-up, which would reduce the need for special fills with intensity con-40 straints. In addition, it would allow to continuously monitor the beam offsets in the collimators with a much better resolution than is currently possible with the standard LHC BPMs, as the distance between the buttons and the beam would be much smaller and there would be no need to interpolate the orbit from the closest BPMs at the collimator location. The collimators could follow 45 orbit drifts and therefore provide more flexibility for local orbit changes, which are regularly required around the experimental insertions. Measuring the beam offset at both ends of the collimator jaws will make it possible to position them fully parallel to the beam trajectory by introducing a longitudinal tilt angle to the jaws. For the time being, the tilt angle of the jaws with respect to the beam 50 can only be evaluated with long and detailed jaw scans and is hence only applied for the injection and dump protection collimators. Furthermore, the margins for orbit drifts between collimator families could possibly be reduced [11], which would eventually allow smaller beam sizes at the experimental interaction points (IPs) and lead to an increased luminosity of collisions.

55
This paper is structured as follows. Section II describes the BLM-based and BPM-based alignment techniques. This is followed by results from multiturn and turn-by-turn beam measurements in Sections III and IV respectively.
Finally, the measurements are compared to simulations using an electromagnetic (EM) model in Section V. 60

BLM-based alignment
The LHC collimators are currently aligned using feedback from the BLMs.
Each jaw is moved separately to the beam on either side until the halo is touched, and the beam center is subsequently calculated as the average of the two aligned left and right jaw positions, J L and J R : (1) Figure 4 shows a typical BLM-based alignment with the mock-up collimator in the SPS. The jaws were moved in steps of 50 µm by means of two stepping motors installed at both extremities. The touching of the beam halo was recorded by a BLM installed about 50 cm downstream of the collimator. One jaw is considered to be aligned, if the signal of the BLM reaches ∼ 1 × 10 −6 Gy/s. This value may vary depending on the average losses without jaw movement, as a spike needs to be clearly distinguished from the background signal. This also defines the minimum step size. Note that in the LHC step sizes of 5 to 20 µm are used due to a better beam quality and higher particle energies. This technique 75 unfortunately does not allow the alignment of the individual jaw corners.

BPM-based alignment
The mock-up collimator consists of two copper jaws and a 10 mm thick Here, X bpm is referred to as the linearized beam position, while X raw is the normalized beam position, calculated from difference over sum of the measured peak voltages V L,R of the opposite electrodes on the left and right jaws: The collimator is aligned when the electrode signals for each jaw corner are  alignment, the gain of the BPM signals was changed as part of the beam test.
At the end, the individual jaw corners are moved until the signals are equalised (or X raw is approximately zero), and the beam center is calculated using Eq. 1.
The electrode signals are proportional to the distance between the beam and 95 the jaw, as well as to the beam intensity, hence one can see a slight decrease in the signals over time.

BPM ELECTRONICS
Experiments with a mock-up collimator with jaw-integrated BPM buttons 100 were performed in the CERN SPS with stored beam at 120 GeV. The beam intensities were usually just below 1 × 10 11 protons (longitudinal length of 4σ = 2 ns), stored in one bunch. During the measurements presented below, the injaw BPMs were connected to the prototype of a high resolution diode-based orbit measurement system, which was developed at CERN for this application.

105
This system was optimized for multi-turn applications. From measurements with BPMs installed in the LHC the achievable resolution with this system was estimated to be well below 1 µm [13].

Measurements with a four corrector closed orbit bump
To compare the accuracy of the BPM-based alignment method with the

Measurements with primary and secondary protons impacting on the jaw
One possible obstacle for the use of collimators with jaw-integrated BPM buttons could be a disturbance of the BPM signals due to particles impacting on the jaw. Therefore, several full beam scrapings with the maximum jaw move-

LHC BPM ELECTRONICS
The use of collimators with in-jaw BPM buttons may also be interesting in the transfer lines between the SPS and the LHC. In this case, the shot-by-shot or respectively the turn-by-turn reproducibility of the measured beam offset is the figure of merit.

180
The measurements presented below were performed with standard LHC BPM electronics connected to the in-jaw BPM buttons in single pass operation. The beam offset in the collimator was recorded in every turn for a total number of 300 turns before the jaws were moved again. of BPMs, introduces gains and offsets in the measured data. Therefore, the BPM data were corrected by applying a linear fit of the form:

Collimator scans with constant gap
to the product B × X raw to obtain values for the fit parameters a and b at each gap and beam offset.   Figure 15 shows schematically the simulation conditions for G = 24 mm and G = 16 mm.

Validation of simulated BPM response by measurements
To validate the simulations, an experiment was performed with the mock-up 235 collimator and circulating beam in the SPS. The jaw gaps were chosen similar to the simulated ones. For each G the sets of 3-5 pre-defined beam offsets were measured. Note that for gaps below 16 mm, large beam offsets of over 60 % of G/2 could not be achieved as these would have caused a scraping of the beam and therefore induced losses.

240
The linearity parameter L f , which is a conversion coefficient between the linearized and the original beam positions, is calculated as:  more, the beam orbit drifted slowly during the experiment, which increased the above mentioned alignment error of the jaws.
Nevertheless, this experiment showed that the geometrical non-linearity of the BPM readings can be well reproduced by simulations and therefore allows 260 its correction for the full range of the jaw motion (2 -60 mm).

CONCLUSION
Collimators with jaw-integrated BPMs promise a drastically reduced set-up time of the LHC collimation system -a few seconds per collimator compared to approximately five minutes achieved with BLM-based alignment -making the 265 alignment less dependent on machine stability, as parasitic monitoring without dedicated fills will be possible. Furthermore, they allow to continuously monitor beam offsets at the collimators and thus improve the passive machine protection.
They permit tighter collimator settings, could help improve the beam cleaning, and possibly allow for a higher luminosity in the experimental interaction points.  The non-linear response of the in-jaw BPM buttons has been simulated as a function of the gap width and the beam offset and compared to measurements with beam. A very good agreement was observed, showing that the discrepancy between simulation and measurement is within 5 %. This discrepancy can eventually be accounted for in the signal processing.

285
Taking into account the results of laboratory measurements, the tests in the SPS, and the LHC collimation experience, it can be concluded that the accuracy of a BPM-based collimator set-up will be significantly better than the current BLM-based method. Furthermore, the measurements have shown that the accuracy of in-jaw BPMs in single pass operation is sufficient for the 290 application in the transfer lines of the LHC. Therefore, the experience and results acquired from this feasibility study endorsed the installation of the new BPM-equipped collimators in the LHC.