Longitudinal Beam Profile Acquisition with Bursts of Triggers in the CERN Proton Synchrotron

CERN (European Organization for Nuclear Research) is upgrading its infrastructure of accelerators. The most widely known accelerator is LHC (Large Hadron Collider). After acceleration process bunches of particles are collided and provide many information. Before the bunches will be injected to LHC they have to be accelerated in a chain of other accelerators. One of them is PS (Proton Synchrotron). It accelerates bunches of particles from 1.4 GeV to 26 GeV and is characterised by many RF manipulations (e.g. bunch splitting). The period of the acceleration operation is called cycle and consists of many turns during which bunches are rotating in a vacuum pipe. The acceleration is performed in RF cavities. In the framework of injector upgrade the beam intensity has to be increased by factor of 2. This can cause instabilities which can prevent obtaining the result. Therefore beam has to be observed in more details to do studies. The purpose of this thesis is to evaluate requirements for new acquisition system for measuring longitudinal beam profile in CERN PS. There exist already one solution called "Tomoscope" which was used as a reference. At first an oscilloscope has been integrated with a multi-trigger device by application which allowed to visualise,store data and manage settings. During operation one can choose many starting points for bursts of triggers. This way one can obtain data of many periods of time (i.e. "tri-split") over one acceleration cycle. There is no more the limitation of only one acquisition during one acceleration cycle. The main limitation of the first approach is transfer of data from the oscilloscope and limited memory space. Both of them are alleviated by streaming data from digitizer. Acquired data was stored in a circular buffer. There is one producer - digitizer and many consumers. The following consumers are implemented: live visualisation, single/continuous storing and bunch length analysis. The system enabled for live observation of many events over many acceleration cycles. The analysis of bunch length turned out to be slow on default 64 bit processors. It requires more signal processing oriented approach. I have used FPGA to implement parallelised bunch length analysis. The work was focused not only on functionality but also on gathering performance requirements, needs of accelerator physicists and checking technological feasibility. As a result of the above mentioned improvements the beam profile data along acceleration cycle is observed easily and in more detail. The bunch length and position analysis is done faster and in more reliable way. The management with data and acquisition settings is more convenient.