Atlas lar Back-End Electronics Installation
|
|  Yüklə 1.15 Mb.
|
ATLAS LAr Back-End Electronics InstallationH.Braun, X. De La Broise, M. Citterio, L. Hervas, T. Hott, L. Kurchaninov, B. Laforge, F. Lanni, B. Lund, R. McPherson, V. Paolone
AbstractThis document describes installation procedures of the back-end electronics for LAr calorimeters. We present the operations with cables in the USA15 hall as well as preparation, installation and test procedures foreseen for LAr back-end systems. Manpower and time issues are also addressed.
The installation of Liquid Argon detectors is basically a sequence of three big activities: (i) cryogenic services and cryostats, (ii) front-end electronics, and (iii) back-end electronics. These three works are performed by different groups of experts and partly separated in time. The installation and commissioning of the front-end electronics in UX15 is described in a separate Note [1] and not addressed here. From the test point of view, this activity is very complex and requires efforts of many experts in many different fields. The back-end system installation requires fewer efforts. The elements in USA15 are standard sub-racks, like VME crates, computers, standard modules like VME boards or transition modules. Nevertheless, the activities in USA15 are very important. The installation of the back-end subsystems has to be properly scheduled because these are the tools for the front-end crates qualification. At the same time, back-end parts have their own production schedule, not always correlated with the schedule of front-end electronics production and installation. In order to reduce dependencies, in the initial stage of the front-end installation, a dedicated small-scale set of cables and back-end parts will be used for the commissioning tests of the front-end crates. This small system is out of scope of this Note. The Liquid argon back-end electronics consists of five subsystems. Table 1 presents different types of sub-racks and modules and Figure 1 shows layout in USA15. In total the Liquid argon back-end system counts up to 149 sub-racks and 819 modules*. This amount of parts will be installed and put to operation during a period of more than a year**, so it cannot be considered as a big amount. The most time consuming operation is not the insertion of units but the testing and debugging procedures. Most of the back-end elements are custom designed hardware, so the participation of the particular experts will be required in the initial phase of installation. In the routine phase, the experts will be needed only in the case of problems. It is assumed that installation team can do the most part of the work. In the case of subsystem-specific actions, like deep debugging, the presence of the physicist or engineer with better knowledge of the subsystem is needed during the period scheduled for such kind of actions. And finally, specific experts are needed if a hardware or a software part is not working properly. These experts are called by installation responsible person and should be able to arrive to CERN within a reasonably short time. The installation sequence of the back-end system can be generally divided to three phases: (i) physical installation, (ii) systematic tests, and (iii) common tests with front-end side. The physical installation includes insertion of sub-racks and modules, cable connections and basic tests to be sure that the parts are operational. The test phase is the full debugging of the subsystem or of its parts. It consists of a set of functional tests and guarantees that the subsystem is ready for use. The last phase, the combined front-end and back-end tests, is a common work of the front-end and back-end groups and is not considered in this Note. The subsystems layout in USA15, the installation sequence, and details on the tests are described in the sections 3-7. Each subsystem will need to make some pre-installation tests in the electronics maintenance facility (EMF). The activities foreseen in the EMF are also addressed in the following sections. In the Summary we present the organization of the EMF test area as it is seen now. Table 1: Sub-racks and modules of the LARG subsystems in USA15.
The other big activity in USA15 is the installation of cables. This work is under shared responsibility of the ATLAS Technical Coordination and Liquid Argon Collaboration [2]. In the next section we describe all those cable operations in USA15 provided by Liquid Argon Collaboration. According to the present ATLAS installation scenario, the barrel cryostat will stay on the truck until June 2005. In this case there is a few month of access to cryostat and front-end installation can be started on-truck position. In this case the back-end electronics will be connected to front-end crates with a dedicated set of cables borrowed from spare parts. If necessary, these cables will be used for initial phase of installation in the final position of barrel. A rough schedule of the back-end installation is considered in the Summary. As it was mentioned, it is driven by the front-end schedule, as well as by the delivery schedule which are not completely defined today. The time needed for different operations is calculated assuming that one working day has 6 pure working hours and one working week is 5 full working days. All the time estimations presented in this Note follow the problemless scenarios. 
All cables connecting LAr detectors and USA15 (except of front-end fibers) are produced with double length and with connectors equipped at both ends. This allows the easy installation in the cable chains and passing through the cable holes between UX15 and USA15. Then cables will be cut and equipped with the second connectors in USA15, after the laying down. The LAr team is responsible to organize and carry out the equipping these second connectors. TC will assume the additional costs generated by the fact that this operation is done in USA15 (instead of a workshop or lab). 2.1. L1 trigger cables The LARG L1 trigger cables deliver analog signals from Tower Builder (Driver) Boards to the Receiver Stations. One cable has typical length of 70 m and is made of 16 shielded twisted pairs with a common protecting shield [3]. In total there are 128 cables for EMB, 112 for EMEC, 96 for HEC and 24 for FCAL that sums up to 360 cables. In addition, there are 14 spare cables placed in the same trays. The trigger cables are produced by a manufacturer but get functionally tested at Saclay. An automatic test bench has been developed for these tests. A signal, similar to Tower Builder Board output, is injected to one end of the cable and measured at the other end. For each pair, signal shape, amplitude and delay are recorded. The crosstalk between pairs will be also evaluated. The absolute gain and its dispersion between pairs, the peaking time and its dispersion, the propagation time and crosstalk will be determined and compared with tolerance values. Out of specification cables will be returned to manufacturer. Saclay controls also the mechanical aspects of the cable, and fills one data sheet per cable. Cables are labeled by the manufacturer at both sides (each side is marked "A" or "B"), so that it is possible to identify the cables after they are cut. The test bench consists of a digital oscilloscope, pulse generator, a CAMAC crate with IEEE controller board and two home-made multiplexer-demultiplexer boards (transmitter and receiver). The PC with LabView software controls the test bench, performs the data analysis, compares numbers with specification values, delivers OK/not-OK message and stores cable characteristics. Cables are tested one by one. This test bench will also be used for the test of the cables after the installation. The way, how to inject signal at the UX15 side is still to be defined. One of the possibilities is to use a dedicated reference cable to deliver signal from the pulser to the cable input. If a cable does not pass the test, different possibilities can be considered. If an excessive crosstalk or amplitude loss is detected, the second connector should be cabled again on the spot. If a mechanical damage is detected, the cable will be repaired if possible. Otherwise one of the two spares per chain can be used. The testing procedure for one cable requires about 30 min. Therefore some 12 cables can be tested within one working day. All cables can be tested within 30 days if time is not lost for the cables re-plugging during the tests. It implies that one person is permanently present in UX15 for manual operations and one or two are working in USA15 site. The test bench and all specification documents will be provided by CEA-Saclay with manpower to perform tests of the EMB and EMEC cables. The testing of HEC and FCAL cables is the responsibility of MPI-Munich and University of Arizona respectively.
The SPAC signals, both upstream and downstream, of each front end crate controller board are transmitted optically from/to their corresponding SPAC Master board located in the read out crate in USA15. The transmission is inhomogeneous in the sense that the upstream links use 850nm wavelength when the downstream lines use 1310 nm wavelength. These signals are transmitted through multimode optical fibers identical to those used to transmit data from FEB to the ROD boards. FEB to ROD links and SPAC links are bundle in the same lot of fibers. For SPAC purposes, the number of fibers per crate is 8 except for the special EMEC-HEC crates where 12 fibers are needed.
The RF timing generators are located at the Prevessin Control Room (PCR) and deliver the constant frequency 40.079 MHz LHC bunch clock, a pseudo LHC orbit signal obtained by dividing the clock by 3564, the real SPS orbit signal and the ramping SPS 40 MHz clock obtained by dividing the SPS RF by 5. The PCR transmitters can each broadcast only one orbit and one clock signal simultaneously. The selected pair are encoded and used to modulate a high power laser, the output from which is split by a 1:32 optical tree coupler and broadcast via optical fiber to different destinations around CERN sites over few kilometers. One of the destinations will be ATLAS experiment. These signals are transmitted using monomode optical fibers with a very small jitter (~7ps). Each experiment is equipped with the TTC machine interface (TTCmi), a part of the Central Trigger system. At that level the signals are duplicated such as to be provided to each partition of the experiment. At the partition level, the trigger and timing signals are transmitted to the LTP module which is interfaced to TTCvi and TTCex modules most frequently coupled to a TTCoc module that fans out the optical signals to the number of desired destinations in the partition. The connections from the TTCex to the TTCoc are optical as the ones from TTCoc to end user boards, namely the FEC controller boards in the case of liquid argon calorimeters partitions. For each controller board there are two optical fibers connected whereas only one is connected on the TTCoc side. One of these two fibers is installed for redundancy and would need to be connected to the TTCoc of the partition only in case if the first connected one fails. In such a case the two fibers would be exchanged manually in USA15. Number of TTC fibers to be installed among the 6 LAr partitions is given in Table 2 for front-end destinations and Table 3 for back-end part. All these fibers will be installed by TDAQ community for all ATLAS partitions. This installation will be done in advance to LAr electronics installation. Table 2: Front End crates of the Liquid Argon calorimeters and TTC links.
Table 3: Read Out crates of the Liquid Argon calorimeters and TTC links.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||