【作者】 向常洲；

【导师】 周代翠； Norbert Herrmann；

【作者基本信息】 华中师范大学 ， 无线电物理， 2013， 博士

【摘要】 在过去的几十年里,得益于高速发展的科学与技术,我们能够更加深入地认识和了解这个自然世界。大型强子对撞机(LHC),作为当今世界上最大并且能量最高的粒子加速器,把人类带入了TeV能级的束流世界。它与相对论重离子对撞机(RHIC)及其它高能对撞机一样,有助于我们认识量子色动力学(QCD)相图。这些加速器及其实验都专注于极高能量和低重子数密度的研究。为了拓展相图中高重子数密度区域的研究,德国在其反质子和离子研究中心启动了压缩重子物质实验(CBM)和SIS100/300加速器的建造。由于实验中粒子的相互作用速率高达10兆赫兹和便于观测各种探针信号,难于设计出一个有效而且高速的触发系统,因此CBM将采用无触发的数据获取系统。作为CBM实验中的核心探测器之一,飞行时间探测器(TOF)主要用来鉴别固定靶实验中产生的带电粒子。在无触发的数据获取系统下,为了得到高达80ps的时间分辨率,设计出高时间分辨率的探测器,高速电子学和读出技术,以及数据处理系统是非常重要。飞行时间探测器计划使用平行板阻抗室。并将采用以PADI (PreAmplifier DIscriminator)芯片和GET4(GSI Event-Driven TDC with4channels)芯片为基础的前端电子学系统。SysCorc主板将实现数据从前端电子学到数据获取系统的传输。由于CBM实验采用无触发的数据获取系统,前端电子学产生的所有各类数据都将被发送到DAQ系统。在不影响探测器性能的情况下,为了尽可能地降低传输的数据量,数据预处理将作为额外的功能模块被设计在飞行时间探测器的在线读出链中。在海德堡大学物理研究所的两年学习期间(2010年10月至2012年10月),我的主要工作就是在读出链中设计并实现数据预处理功能。在该论文中,我将主要阐述基于GET4芯片数据的分析,设计并测试TOF读出链上的数据预处理功能。论文的前两章,主要介绍CBM实验和TOF探测器。其中,包括两种TOF探测器候选方案的片状RPC和条状RPC。数据预处理的设计与RPC探测器的几何有关。第三章主要介绍当前TOF探测器前端电子学及其读出系统,作为数据预处理的设计基础,将详细介绍PADI板和GET4板。另外,由于前端电子学及其读出系统仍处于研发阶段,在章节的最后将介绍它们的最新结果。由于数据预处理与前端电子学及其读出的硬件密切相关,分析和评估具有数据预处理功能模块的性能是非常必须的。第四章首先介绍一种典型的离线分析系统是如何分析GET4数据的。该典型离线分析系统主要用来分析基于"PADI+GET4+ROC"的数据链。其次,本章描述了当前读出系统的ROC数据速率和噪声与PADI阈值的关系。束流测试的分析结果表明,GET4数据的时间排序问题,降低数据存储量和过滤噪声需要在数据被发送至DAQ之前解决。第五章给出了数据预处理模块。数据预处理模块包括三个单元：数据预处理单元,监控单元和控制单元。数据预处理单元作为主要部分,主要用来实现两个功能：粒子过滤和簇过滤。粒子过滤的目标是去除无用的GET4数据,确定粒子到达TOF探测器的时间和能量,降低传输到DAQ的数据量。簇过滤用来实现相关性粒子的收集和去除无用的粒子信息数据。另外,簇过滤为CBM实验第一级事件筛选First Level Event Selection (FLES)做准备。监控单元用来实现对前端电子学和ROC主板的监控。控制单元用来完成对PADI芯片阈值的自动设定,由于受限于ROC主板上主芯片的资源,这部分并未在数据读出链中设计并实现。在将数据预处理模块植入到读出链中之前,需要确定相应的参数并对模块的性能进行评估。为此,设计出基于Go4(GSI Object Oriented On-line-Offline)的C/C++程序。该程序能够实现离线数据分析和进行数据获取时的在线监控。特别重要的是,它在处理GET4数据时,采用了与ROC主板上主控芯片FPGA相同的方式。利用该程序和已有的GET4原始数据,进行了功能仿真。通过仿真,得到了正确的参数,并分析出了数据预处理模块的性能。在2011年,在德国于利希研究中心,利用束流测试了具有粒子过滤功能的数据预处理模块。在2012年,在海德堡大学物理研究所实验室,测试了具有粒子过滤和簇过滤功能的数据预处理模块。测试平台的搭建和结果见第六章。综上所述,论文的主要内容如下,·在读出控制板(ROC)固件中设计并实现了在线数据预处理模块,束流测试的结果表明,在线数据压缩率能够达到90%以上。·在离线分析软件中开发了针对于TOF的数据流软件,并将软件移植到无触发的在线环境中。·对整个CBM-TOF探测器的数据预处理链提出了目标要求,在当前读出控制板(ROC)固件中实现了相应的接口。由于前端电子学,读出电子学,数据获取系统和FLES系统都在进一步的研发之中,为CBM-TOF探测器设计的数据预处理模块工作仍在继续,本论文中涉及的研究工作仅仅只是一个开始。更多还原

【Abstract】 In the last decades, our nature was deeper understood benefiting from the fast development of science and technology. The Large Hadron Collider (LHC), which is the world’s largest and highest-energy particle accelerator, brings human to enter TeV beam energy level. Like Relativistic Heavy Ion Collider (RHIC) and other high energy colliders, it helps us explore the Quantum Chromo-Dynamics (QCD) phase diagram of strongly interacting matter. These accelerators and experiments are focused on the very high energy with low baryonic density. To exploring the phase diagram in the region of very high baryonic densities, the Compressed Baryonic Matter (CBM) experiment and SIS100/300of the Facility for Antiproton and Ion Research (FAIR) are planned in Germany. As the high interaction rate is up to10MHz for the Au+Au collisions at25AGeV, an efficient fast trigger is difficult to be implemented. Thus, a free stream data acquisition (DAQ) is planned for the CBM. The Time-of-Flight (TOF) detector is one of the core detectors of the CBM experiment. It provides particle identification for all charged hadrons emitted from the beam-target interactions.In order to achieve a time resolution of80ps within the free stream DAQ system, it is important to design high time resolution detector, with fast readout electronics, and online data processing system. The TOF detector is decided to use Resistive Plate Chamber (RPC) for the whole wall. The front-end electronics with PADI (PreAm-plifier DIscriminator) chip and GET4(GSI Event-Driven TDC with4channels) chip is an option for the detector. The SysCore board is used as Readout Control (ROC) for data transferring to DAQ. All kinds of data will be readout from front-end elec-tronics and sent to DAQ for the CBM experiment operating a free streaming DAQ system. In order to reduce the bandwidth without detector performance loss, an additional function, data pre-processing, is required to be implemented in the TOF readout chain. The main work during my stay at Physikalisches Institut Heidelberg from October2010to October2012is to implement this data pre-processing in the readout chain. In this thesis, I will present the analysis of GET4chip data, design and test of data pre-processing for TOF readout chain.In the first two chapters, the CBM experiment and the TOF detector are intro-duced. Both the pad RPC and strip RPC are considered as for the TOF detector, and are described in detail for the data pre-processing is related to the geometry of the RPC detectors.Chapter3introduces the current (during my stay in Heidelberg) TOF front-end electronics and its readout. The PADI board and GET4board are described in detail. The design of data pre-processing is based on the characters of these boards. As the front-end electronics and its readout are still under developed, the new version of them is described at the end of the chapter, although they are not considered in the design of data pre-processing.As the data pre-processing is very specific to the front-end electronics and readout, the analysis and evaluation on the data pre-processing is necessary. In chapter4, a typical offline analysis on GET4data is described. The analysis is a preliminary standard way for the PADI+GET4+ROC readout chain. The ROC data rate based on the current readout chain is illustrated in this chapter and also the relation between noise rate and PADI threshold. The result from a beam test is presented. It shows that the time ordering of GET4data, data storage reduction and filtering noise should be solved before being sent to DAQ.Referred on the analysis in chapter4, we designed a data pre-processing for ROC firmware in chapter5. It is consists of three unit, a data pre-processing unit, a monitor unit and a control unit. The data pre-processing unit is the main part. It has two functions, hit filtering and cluster filtering. The goal of hit filtering is to remove useless data from GET4chip and to reduce the data rate to DAQ by transferring hits with time stamp and time-over-thrcshold (TOT) values instead of edge data from GET4chip. The cluster filtering focuses on collecting relative hits and removing useless hits. Further more, the cluster filtering is the first stage of data processing for First Level Event Selection (FLES). The monitor unit is designed to monitor the front-end electronics and ROC itself. The control unit is planned to realize the automatic threshold setting. Due to the limited resource of current FPGA used in the ROC, it has not been implemented yet.Before the data pre-processing to be implemented in the readout chain, its pa-rameters have to be found out and its performance has to be evaluated. A C/C++code for these tasks is designed under the Go4(GSI Object Oriented On-line-Offline) environment. The code can run for offline analysis and online monitoring. Specially, it uses the method as in FPGA to process the GET4data. A simulation with raw data got from a beam test as inputs is done with the code and correct parameters are determined during the simulation. The data pre-processing with the hit filtering function implemented was tested in beam in November2011in Julich in Germany. The data pre-processing with hit filtering and cluster filtering functions was tested in lab. The test setup and results are presented in chapter6.The main tasks and contributions are as followed.·Implement module on the data pre-processing into the readout controller (ROC) firmware and realized online data suppression for testing beam data of more than90%.· Developed the software for expecting streaming TOF data in offline analysis framework and transfer the codes to free streaming online environment.· Identified the main requirement for the data preprocessing of the full CBM-TOF data processing chain and implemented interface in the current ROC firmware. As the front-end electronics, readout electronics, DAQ and FLES system are still in development, the data pre-processing functions for CBM-TOF in readout chain will enter developing stage in future.更多还原