Academia.eduAcademia.edu

Compressed Baryonic Matter experiment at FAIR - Progress Report 2018

Profile image of Ryszard RomaniukRyszard Romaniuk

2019, Compressed Baryonic Matter experiment at FAIR - Progress Report 2018

https://doi.org/10.15120/GSI-2019-01018
visibility

…

description

220 pages

link

1 file

Abstract

An experimental project like CBM necessarily comprises a vast number of activities in many different areas: research on detector technology, development of readout electronics and components for data acquisition, computing and software tools for data processing and physics analysis, and many more. The series of annual CBM Progress Reports, started back in 2006, was intended to collect and document these manifold activities. Browsing through the past volumes unfolds a large spectrum of scientific work in the process of the realization of the project: fromconceptual studies over thoroughR&Dto the implementation and testing of prototypes. This CBM Progress Report 2018 continues along these lines. Its contents, however, reflect that some six years before the planned start of data taking, the CBM project is undergoing a gradual transition. The long period of planning and R&D is giving way to the large-scale production and integration of detector hardware, a process to be finished by 2024, when the CBM apparatus is expected to be commissioned in its experimental area. An important step towards this realisation of the experiment is the full-system test setup mCBM, allowing to study the joint in-beam operation of several detector systems and the read-out and data processing, following the ambitious CBM concept of free-running data acquisition. Further important technological experience is gained by the deployment of CBM detector systems at running experiments: TOF in STAR, RICH in HADES, PSD in BM@N. These detector operations will contribute important technological expertise for the preparation of the full CBM experiment. We hope this reports conveys some of the enthusiasm of the CBM collaboration in the realization of a technologically very challenging experimental project which promises a rich physics output once taking data. Our thanks go to all who have contributed to this report: the reviewers, who helped getting it into shape, and all authors having delivered the actual content. Darmstadt, October 2019 Volker Friese and Ilya Selyuzhenkov, editors

Figures (395)

The exploration of the QCD phase diagram in the regio! of high baryon densities is the primary goal of the physic program of the Compressed Baryonic Matter (CBM) expe iment at FAIR. During 2018 the collaboration faced a de tailed evaluation in the context of the FAIR project revie that was successfully passed with the arguments presente below. Figure 1: Rate capabilities as function of collision energy of existing and planned experiments (Figure adapted from [1]).
The exploration of the QCD phase diagram in the regio! of high baryon densities is the primary goal of the physic program of the Compressed Baryonic Matter (CBM) expe iment at FAIR. During 2018 the collaboration faced a de tailed evaluation in the context of the FAIR project revie that was successfully passed with the arguments presente below. Figure 1: Rate capabilities as function of collision energy of existing and planned experiments (Figure adapted from [1]).
Figure 2: CBM setups for different physics topics: Hadron (left), Electron/Hadron(center) and Muon(right). For de- tails see text.
Figure 2: CBM setups for different physics topics: Hadron (left), Electron/Hadron(center) and Muon(right). For de- tails see text.
Figure 1: Accumulated hits from an °°Fe radioactive source as recorded with the analogue output of MIMOSIS- 0. The signal of the on-pixel pre-amplifier (upper signal) and of the on-pixel discriminator (inverted lower line) are shown.
Figure 1: Accumulated hits from an °°Fe radioactive source as recorded with the analogue output of MIMOSIS- 0. The signal of the on-pixel pre-amplifier (upper signal) and of the on-pixel discriminator (inverted lower line) are shown.
Figure 1: Average hit number per frame vs. sensor ID for an Au+Au (upper) in comparison to a p+Au reaction, shown for all four station half sides, assuming left-right stations geometry. In conclusion, the results presented and detailed in [1] pro- vide firm estimates for planning the number of GBTx chips employed by MVD, as well as e.g. designing the layout of the electrical feed-throughs of the target chamber front plate. They are based on the projected MIMOSIS r/o ar- chitecure and geometrical dimension, considering the nom- inal interaction rates for AA and pA reactions and conceiv- ably reduced mgn. field.
Figure 1: Average hit number per frame vs. sensor ID for an Au+Au (upper) in comparison to a p+Au reaction, shown for all four station half sides, assuming left-right stations geometry. In conclusion, the results presented and detailed in [1] pro- vide firm estimates for planning the number of GBTx chips employed by MVD, as well as e.g. designing the layout of the electrical feed-throughs of the target chamber front plate. They are based on the projected MIMOSIS r/o ar- chitecure and geometrical dimension, considering the nom- inal interaction rates for AA and pA reactions and conceiv- ably reduced mgn. field.
Figure 2: CAD illustrative view on the keep-out volumes located outside (A) and inside (B) the target chamber, with void areas (V1 - target volume, V2 - detector support vol- ume, V3 - vacuum window between MVD and STS). Figure 1: Schematic side (left) and front (right) view of the MVD setup, including support structures and functional unites referred to in the text; (0,0,0): target position.
Figure 2: CAD illustrative view on the keep-out volumes located outside (A) and inside (B) the target chamber, with void areas (V1 - target volume, V2 - detector support vol- ume, V3 - vacuum window between MVD and STS). Figure 1: Schematic side (left) and front (right) view of the MVD setup, including support structures and functional unites referred to in the text; (0,0,0): target position.
The MVD [1] is the vertex silicon pixel detector of the CBM experiment, placed closest to the target and in front of the STS. Based on our current knowledge, we have been working on a realistic MVD CAD model that could help in joining forces resulting in a common understanding of the MVD environment, see Figure 1, the target volume and the installation process.
The MVD [1] is the vertex silicon pixel detector of the CBM experiment, placed closest to the target and in front of the STS. Based on our current knowledge, we have been working on a realistic MVD CAD model that could help in joining forces resulting in a common understanding of the MVD environment, see Figure 1, the target volume and the installation process.
Figure 1: PRESTO (1) mounted on movable table (2); FEE board (3) interconnecting low-mass flex (orange) with stan- dard flex (white) cables; front flange (4), TRBv3 r/o (5), customized feed-through flange (6) for r/o cables. Figure 2: Trend plots of environmental parameters recorded 24/7 within a time period of 3 months with PRESTO inside the vacuum vessel.
Figure 1: PRESTO (1) mounted on movable table (2); FEE board (3) interconnecting low-mass flex (orange) with stan- dard flex (white) cables; front flange (4), TRBv3 r/o (5), customized feed-through flange (6) for r/o cables. Figure 2: Trend plots of environmental parameters recorded 24/7 within a time period of 3 months with PRESTO inside the vacuum vessel.
tests and efforts will continue in 2019.
tests and efforts will continue in 2019.
Figure 2: Prototype STS module demonstrating double- sided read-out of a microstrip sensor through 25 cm long microcables with two STS-XYTERv2.0 ASIC boards. Figure 1: Test beam telescope of two scintillating fiber ho- doscopes and two prototype STS modules under study.
Figure 2: Prototype STS module demonstrating double- sided read-out of a microstrip sensor through 25 cm long microcables with two STS-XYTERv2.0 ASIC boards. Figure 1: Test beam telescope of two scintillating fiber ho- doscopes and two prototype STS modules under study.
'GSI, Darmstadt, Germany; 7FAIR, Darmstadt, Germany; *Universitit Tiibingen, Germany; “Goethe Universitit, Frankfurt, Germany; SKINR, Kiev, Ukraine; now Université catholique, Louvain, Belgium; KIT, Karlsruhe, German’
'GSI, Darmstadt, Germany; 7FAIR, Darmstadt, Germany; *Universitit Tiibingen, Germany; “Goethe Universitit, Frankfurt, Germany; SKINR, Kiev, Ukraine; now Université catholique, Louvain, Belgium; KIT, Karlsruhe, German’
Table 1: Global parameters of the tested CBM sensors for mSTS. C=CiS, H=Hamamatsu. The difference in Cy is due to different thickness of the silicon wafers used by CiS and Hamamatsu. One sensor exhibits a breakdown below 200 V of bias voltage while it is required for a sensor to operate without a breakdown up to 200 V. In order to investigate such behavior, IV- CV measurements with guard rings were performed. Fig- ure 1 shows measured IV curves revealing the origin of the breakdown. It occurs through the sensor edge region where defects caused by blade dicing and handling appear. It is considered the most frequent cause of the breakdown of the CiS prototypes. The bulk of all tested sensors does not show any breakdown signs below 500 V of bias volt- age. The total leakage current of the Hamamatsu and CiS sensors is around 7 nA/cm? and 170 nA/cm? at 150 V, respectively. Only sensor 352151-13 has higher current of 250 nA/em? due to a cluster of leaky strips on the p side of the sensor. No visual damage was found for those strips suggesting a manufacturing defect during implantation or a crystal defect.
Table 1: Global parameters of the tested CBM sensors for mSTS. C=CiS, H=Hamamatsu. The difference in Cy is due to different thickness of the silicon wafers used by CiS and Hamamatsu. One sensor exhibits a breakdown below 200 V of bias voltage while it is required for a sensor to operate without a breakdown up to 200 V. In order to investigate such behavior, IV- CV measurements with guard rings were performed. Fig- ure 1 shows measured IV curves revealing the origin of the breakdown. It occurs through the sensor edge region where defects caused by blade dicing and handling appear. It is considered the most frequent cause of the breakdown of the CiS prototypes. The bulk of all tested sensors does not show any breakdown signs below 500 V of bias volt- age. The total leakage current of the Hamamatsu and CiS sensors is around 7 nA/cm? and 170 nA/cm? at 150 V, respectively. Only sensor 352151-13 has higher current of 250 nA/em? due to a cluster of leaky strips on the p side of the sensor. No visual damage was found for those strips suggesting a manufacturing defect during implantation or a crystal defect.
Table 2: Statistics on the identified defects of the tested CBM sensors for mSTS. The total amount of bad strips per sensor per side is well below the required acceptance limit of 1.5% for the CBM- STS [2] except for the sensor with ID number 5552-301 where all 27 bad strips belong to the p side of the sen- sor. However, the sensor is fully operational and still can be used for module assembly in mSTS. Because implant
Table 2: Statistics on the identified defects of the tested CBM sensors for mSTS. The total amount of bad strips per sensor per side is well below the required acceptance limit of 1.5% for the CBM- STS [2] except for the sensor with ID number 5552-301 where all 27 bad strips belong to the p side of the sen- sor. However, the sensor is fully operational and still can be used for module assembly in mSTS. Because implant
Figure 1: Bulk and edge currents of a 6 x 6 cm? CiS sensor as a function of reverse bias voltage. "University of Tiibingen, Germany; 2KINR, Kiev, Ukraine; *GSI, Darmstadt, German
Figure 1: Bulk and edge currents of a 6 x 6 cm? CiS sensor as a function of reverse bias voltage. "University of Tiibingen, Germany; 2KINR, Kiev, Ukraine; *GSI, Darmstadt, German
Figure 2: Example of the results of the strip integrity scan on one of the CBMO06 sensors: Strip series capacitance (top) and dielectric current at 20 V across the coupling ca- pacitors (bottom).
Figure 2: Example of the results of the strip integrity scan on one of the CBMO06 sensors: Strip series capacitance (top) and dielectric current at 20 V across the coupling ca- pacitors (bottom).
Figure 3: Cluster of strips with broken implants. Dark areas are interstrip gaps, horizontal lines correspond to second metal layer. Implant breaks can be seen through the thin metalization layers shown in white.
Figure 3: Cluster of strips with broken implants. Dark areas are interstrip gaps, horizontal lines correspond to second metal layer. Implant breaks can be seen through the thin metalization layers shown in white.
A single sensor of the Silicon Tracking System (STS) is connected via thin microcables to two FEB8 electron- ics boards (n-side and p-side). Each FEB8 PCB can be equipped with eight STS-XYTER readout ASICs. Figure 1: Previous version of the FEB8 board.
A single sensor of the Silicon Tracking System (STS) is connected via thin microcables to two FEB8 electron- ics boards (n-side and p-side). Each FEB8 PCB can be equipped with eight STS-XYTER readout ASICs. Figure 1: Previous version of the FEB8 board.
Figure 2: New FEB8 board layout with test points.
Figure 2: New FEB8 board layout with test points.
Figure 3: APS adapter for PCB from WinWay Technology.
Figure 3: APS adapter for PCB from WinWay Technology.
Figure 1: Tracking station 0 of the mSTS, made from Unit 0 Cleft) and Unit 1 (right) shown in an unfolded view. Every Unit has one ladder with two modules mounted. The sensors are shown in blue along side the module type code. For the first time, real STS modules have been assem- bled for the first tracking station of the mSTS detector. The station consists of two units, each with one ladder that is equipped with two sensor-modules (Fig. 1). Based on the experience with dummy modules the assembly was per- formed according to the detailed process workflow sheet that documents every step taken. Silicon sensors were cho- sen with regard to their characteristics, the STS-XYTER ASICs were tested and calibrated before use and also the quality of the TAB- and wirebonding steps was checked during the assembly.
Figure 1: Tracking station 0 of the mSTS, made from Unit 0 Cleft) and Unit 1 (right) shown in an unfolded view. Every Unit has one ladder with two modules mounted. The sensors are shown in blue along side the module type code. For the first time, real STS modules have been assem- bled for the first tracking station of the mSTS detector. The station consists of two units, each with one ladder that is equipped with two sensor-modules (Fig. 1). Based on the experience with dummy modules the assembly was per- formed according to the detailed process workflow sheet that documents every step taken. Silicon sensors were cho- sen with regard to their characteristics, the STS-XYTER ASICs were tested and calibrated before use and also the quality of the TAB- and wirebonding steps was checked during the assembly.
suring the noise level. Good ASICs were chosen for as- sembly and tests were implemented after the TAB-bonding process of microcables to ASICs as can be seen in Fig. 2. The biggest advantage during this stage is the possibility of rebonding any unconnected channel found.
suring the noise level. Good ASICs were chosen for as- sembly and tests were implemented after the TAB-bonding process of microcables to ASICs as can be seen in Fig. 2. The biggest advantage during this stage is the possibility of rebonding any unconnected channel found.
Figure 3: STS-XYTER test socket on the bonding machine. As a third stage, ASICs were tested in order to check for the electrical connectivity between the ASIC, microca- ble and sensor with same procedure as in the second stage. Likewise, during this stage it is possible to rebond any un- connected channel at the sensor side. All test stages took 5 minutes per ASIC. The tests were performed on the bond- ing machine as shown in Fig. 3. Unconnected channels were identified by checking the hit distributions in the noise pattern. As an example, Fig. 4 shows the connection check result of an ASIC. The uncon- nected channel, seen as a continuous line at low channel number, becomes visible when evaluating the hit distribu- tion in terms of ADC values.
Figure 3: STS-XYTER test socket on the bonding machine. As a third stage, ASICs were tested in order to check for the electrical connectivity between the ASIC, microca- ble and sensor with same procedure as in the second stage. Likewise, during this stage it is possible to rebond any un- connected channel at the sensor side. All test stages took 5 minutes per ASIC. The tests were performed on the bond- ing machine as shown in Fig. 3. Unconnected channels were identified by checking the hit distributions in the noise pattern. As an example, Fig. 4 shows the connection check result of an ASIC. The uncon- nected channel, seen as a continuous line at low channel number, becomes visible when evaluating the hit distribu- tion in terms of ADC values.
Table 1: Quality statistics of the ASIC bondings. Measurement results from the mSTS module assembly are shown in Table 1. For the module 02Tr untested ASICs were used due to unavailability of the pogopin socket at that time, and for module 01TI only one side could be fully tested. Since some ASICs showed problems on Front End Boards during module operation after the bonding process, it is thought that testing ASICs for several times can lead to damage of the bonding area. Consequently, carrying out bare ASIC tests was decided to be sufficient during module assembly. Thus forthcoming module productions will be done with one ASIC test stage only.
Table 1: Quality statistics of the ASIC bondings. Measurement results from the mSTS module assembly are shown in Table 1. For the module 02Tr untested ASICs were used due to unavailability of the pogopin socket at that time, and for module 01TI only one side could be fully tested. Since some ASICs showed problems on Front End Boards during module operation after the bonding process, it is thought that testing ASICs for several times can lead to damage of the bonding area. Consequently, carrying out bare ASIC tests was decided to be sufficient during module assembly. Thus forthcoming module productions will be done with one ASIC test stage only.
Figure 4: Noise hits read at the ASIC’s ADC counters.
Figure 4: Noise hits read at the ASIC’s ADC counters.
to the FEB8. Finally the module can then be completely tested in a dedicated test stand before it is ready to be inte- grated on a ladder. Figure 5: Testing of the FEB8 after the wire bonding of ASIC-rows.
to the FEB8. Finally the module can then be completely tested in a dedicated test stand before it is ready to be inte- grated on a ladder. Figure 5: Testing of the FEB8 after the wire bonding of ASIC-rows.
Figure 6: Full STS module assembled, awaiting to be fitted with the external microcable shield layer.
Figure 6: Full STS module assembled, awaiting to be fitted with the external microcable shield layer.
Components of the sensor module together with the pogo pin device are being placed inside the shielding box during the measurement (Fig. 1) to avoid the influence of electro- magnetic interference. Figure 1: The bonding tool with a sensor, installed inside the shielding box. The tested microcable plus ASIC goes to the pogo pin device on the left.
Components of the sensor module together with the pogo pin device are being placed inside the shielding box during the measurement (Fig. 1) to avoid the influence of electro- magnetic interference. Figure 1: The bonding tool with a sensor, installed inside the shielding box. The tested microcable plus ASIC goes to the pogo pin device on the left.
The measurement takes a few minutes depending on the echonological step. The states of each electronics channel f the XYTER are shown on the GUI in colors indicating ‘ood and bad bonding, noisy and dead channels (Fig. 2). A istogram of measured noise versus the channel number is lisplayed together with similar histograms measured at the revious technological steps. The results of the measure- nents are stored in ROOT files. Figure 2: GUI: bonding of a microcable to 6x2cm sensor.
The measurement takes a few minutes depending on the echonological step. The states of each electronics channel f the XYTER are shown on the GUI in colors indicating ‘ood and bad bonding, noisy and dead channels (Fig. 2). A istogram of measured noise versus the channel number is lisplayed together with similar histograms measured at the revious technological steps. The results of the measure- nents are stored in ROOT files. Figure 2: GUI: bonding of a microcable to 6x2cm sensor.
A typical gold stud bump-bonding interconnection is not possible for STS, as the gold studs cannot be placed on the compressible microcable. Instead, the gold stud is replaced with fine-grained solder paste which is stencil-printed di- rectly onto the Cu microcable. The individual process steps of the novel bump bonding assembly process based on the Cu microcable are depicted in Fig. 1. Figure 1: Individual process steps of gold stud to solder bump bonding connection.
A typical gold stud bump-bonding interconnection is not possible for STS, as the gold studs cannot be placed on the compressible microcable. Instead, the gold stud is replaced with fine-grained solder paste which is stencil-printed di- rectly onto the Cu microcable. The individual process steps of the novel bump bonding assembly process based on the Cu microcable are depicted in Fig. 1. Figure 1: Individual process steps of gold stud to solder bump bonding connection.
Figure 2: Gold stud bumps on STS-XYTER.
Figure 2: Gold stud bumps on STS-XYTER.
Figure 3: Top: Solder paste printed on Cu microcable. Bottom: Solder paste after reflow. cables are then immediately reflown for ease of handling and storage (Fig. 3). Next, cable and die are connected in a flip chip thermocompression bonding process. A Finetech femto flip chip bonder guarantees an alignment accuracy of 0.5 wm. Finally, Polytec EP 601-LV underfill glue is applied to strengthen and protect the interconnection.
Figure 3: Top: Solder paste printed on Cu microcable. Bottom: Solder paste after reflow. cables are then immediately reflown for ease of handling and storage (Fig. 3). Next, cable and die are connected in a flip chip thermocompression bonding process. A Finetech femto flip chip bonder guarantees an alignment accuracy of 0.5 wm. Finally, Polytec EP 601-LV underfill glue is applied to strengthen and protect the interconnection.
the commercial flip chip machines are not suitable for han- dling the sensor-microcable interconnection, an in-house bonding machine has been developed which is close to fi- nalization. Figure 5: Cross-section of the gold stud to solder paste in- terconnection showing excellent wetting of the gold stud.
the commercial flip chip machines are not suitable for han- dling the sensor-microcable interconnection, an in-house bonding machine has been developed which is close to fi- nalization. Figure 5: Cross-section of the gold stud to solder paste in- terconnection showing excellent wetting of the gold stud.
Figure 6: Cu microcable with liquid resist solder stop showing excellent release of the bond pads. Several versions of test cables with a length of 200 mm and 186.5 mm have been produced. Special focus was put on the solder stop, as early versions have shown improper solder stop mask development resulting in an unwanted sol- der height of 20 wm. The latest microcable version is based on a liquid resist solder stop. The pad release is excellent as can be seen in Fig. 6.
Figure 6: Cu microcable with liquid resist solder stop showing excellent release of the bond pads. Several versions of test cables with a length of 200 mm and 186.5 mm have been produced. Special focus was put on the solder stop, as early versions have shown improper solder stop mask development resulting in an unwanted sol- der height of 20 wm. The latest microcable version is based on a liquid resist solder stop. The pad release is excellent as can be seen in Fig. 6.
Figure 4: Cu microcable (at the lower edge of the photo) connected to STS-XYTER with gold stud to solder paste bump bonding technology.
Figure 4: Cu microcable (at the lower edge of the photo) connected to STS-XYTER with gold stud to solder paste bump bonding technology.
Figure 1: Schematic picture of module connectivity. Within the model of the STS, the “Base Length” (Lg) of the microcables is introduced to systematize the estima- tions. Dp is defined as the distance from the sensor edge to the corresponding cooling fin which is planned for the housing of two FEBs of the module as demonstrated in Fig. 1. The P and N sides of the modules are read out with corresponding layers of the microcable of different lengths schematically depicted by Fig. 2. The Base Length of the microcable can be calculated with the following equation:
Figure 1: Schematic picture of module connectivity. Within the model of the STS, the “Base Length” (Lg) of the microcables is introduced to systematize the estima- tions. Dp is defined as the distance from the sensor edge to the corresponding cooling fin which is planned for the housing of two FEBs of the module as demonstrated in Fig. 1. The P and N sides of the modules are read out with corresponding layers of the microcable of different lengths schematically depicted by Fig. 2. The Base Length of the microcable can be calculated with the following equation:
Figure 3: Base length and number of analog cables. Figure 2: Difference in N and P- layers of the cable. References
Figure 3: Base length and number of analog cables. Figure 2: Difference in N and P- layers of the cable. References
The base lengths for all BM@N STS modules were finally tabulated (Fig. 3) and distributed within the BMN@STS collaborating teams for final approval and sub- sequent production of first batches of cables.
The base lengths for all BM@N STS modules were finally tabulated (Fig. 3) and distributed within the BMN@STS collaborating teams for final approval and sub- sequent production of first batches of cables.
The assembly of the first detector modules is a crucial step towards the realization of the STS. In the framework of the mCBM project, the STS will contribute with two track- ing stations using a total of 13 detector modules. These are built with 6.2 x 6.2 cm? sensors and micro cables of approximately 45 cm length. Every sensor is readout by two prototype frontend boards witch 8 STS-XYTER ASICs each (FEB8). Further details on the assembly and quality assurance are given in [1]. Figure 1: Fully assembled module before its installation in the mSTS setup.
The assembly of the first detector modules is a crucial step towards the realization of the STS. In the framework of the mCBM project, the STS will contribute with two track- ing stations using a total of 13 detector modules. These are built with 6.2 x 6.2 cm? sensors and micro cables of approximately 45 cm length. Every sensor is readout by two prototype frontend boards witch 8 STS-XYTER ASICs each (FEB8). Further details on the assembly and quality assurance are given in [1]. Figure 1: Fully assembled module before its installation in the mSTS setup.
Figure 2: First tests of a fully assembled module before its installation in the mSTS setup. Top-panel shows the estimated noise level per channel. In the bottom panel is shown the distribution of the ADC gain for both polarities. 'Goethe University, Frankfurt, Germany; *GSI, Darmstadt, Germany; *Istanbul University, Istanbul, Turkey; 4Tiibingen University, Tiibingen, Germany
Figure 2: First tests of a fully assembled module before its installation in the mSTS setup. Top-panel shows the estimated noise level per channel. In the bottom panel is shown the distribution of the ADC gain for both polarities. 'Goethe University, Frankfurt, Germany; *GSI, Darmstadt, Germany; *Istanbul University, Istanbul, Turkey; 4Tiibingen University, Tiibingen, Germany
Figure 1: Test bench (top). Test station with baby sensor (bottom).
Figure 1: Test bench (top). Test station with baby sensor (bottom).
Figure 3: Energy deposition Station 1 N_ side, 150 MeV (top); 50 MeV (bottom). Figure 2: Beam profile, Station 1 (top); Station 2 (bottom).
Figure 3: Energy deposition Station 1 N_ side, 150 MeV (top); 50 MeV (bottom). Figure 2: Beam profile, Station 1 (top); Station 2 (bottom).
Figure 4: Time difference between signals from N and P side of Station | (top); Station 2 (bottom).
Figure 4: Time difference between signals from N and P side of Station | (top); Station 2 (bottom).
Table 1: Specifications of the pre-series ladders.
Table 1: Specifications of the pre-series ladders.
Figure 1: Carbon fiber ladder on a hub (photo: ICM).
Figure 1: Carbon fiber ladder on a hub (photo: ICM).
Figure 2: Pre-series carbon fiber ladders of 1.2 m length.
Figure 2: Pre-series carbon fiber ladders of 1.2 m length.
Figure 1: (a) Positioning tool on the base plate; (b) Module positioned onto the ladder. A large scale prototype tool (Fig. 1 (a)) has been de- signed to assemble STS ladders. Initially, a half-ladder has been populated with 5 non-functional modules to prove the assembly procedure. The tool has further been used to assemble a ladder with functional modules for use in the mCBM demonstrator experiment. Using the tools, the mounting precision of sensors is limited by the mechanical tolerances of the tools and fittings.
Figure 1: (a) Positioning tool on the base plate; (b) Module positioned onto the ladder. A large scale prototype tool (Fig. 1 (a)) has been de- signed to assemble STS ladders. Initially, a half-ladder has been populated with 5 non-functional modules to prove the assembly procedure. The tool has further been used to assemble a ladder with functional modules for use in the mCBM demonstrator experiment. Using the tools, the mounting precision of sensors is limited by the mechanical tolerances of the tools and fittings.
Figure 2: (a) Module positioned onto the ladder; (b) Proto- type ladder assembled with five non-functional modules.
Figure 2: (a) Module positioned onto the ladder; (b) Proto- type ladder assembled with five non-functional modules.
Figure 3: (a) Histogram showing the deviation of the mea- sured Z-position from the nominal position; (b) Measure- ment of the sensor surface space points with a step size of 5 mm. The black dots refer to the alignment marks on the sensors.
Figure 3: (a) Histogram showing the deviation of the mea- sured Z-position from the nominal position; (b) Measure- ment of the sensor surface space points with a step size of 5 mm. The black dots refer to the alignment marks on the sensors.
Figure 4: First functional ladder assembled for the mSTS detector in the assembly fixture. Two modules with one 6.2 by 6.2 cm sensor each are mounted. The sensors, located on the left hand side of the carbon frame, are covered by the microcable shielding layer that was not removed there yet.
Figure 4: First functional ladder assembled for the mSTS detector in the assembly fixture. Two modules with one 6.2 by 6.2 cm sensor each are mounted. The sensors, located on the left hand side of the carbon frame, are covered by the microcable shielding layer that was not removed there yet.
Figure 1: (a) L-legs glued on the surface of the sensor with 5 different adhesives; (b) L-legs remaining attached to the sensor’s surface after the pull test.
Figure 1: (a) L-legs glued on the surface of the sensor with 5 different adhesives; (b) L-legs remaining attached to the sensor’s surface after the pull test.
Figure 2: Pull-off forces for different adhesives.
Figure 2: Pull-off forces for different adhesives.
Three PCBs with 4 LVRs each had been delivered from Semi-Conductor Laboratory Chandigarh/India [1] for mea- surements and irradiation tests. The LVRs on the boards were supplied pairwise with V;,, and the load on the output side was chosen such that at V.,,, = 1.8 V the output current was set to around 1.4 A. The Enable pin was set to GND in all cases so that chips were active and could be observed continuously (in Fig. 1 for a pair of ASICs). Figure 1: Thermal image of one of the three PCBs under test with only two LVR chips active. The resistor traces reached more than 60°C without cooling.
Three PCBs with 4 LVRs each had been delivered from Semi-Conductor Laboratory Chandigarh/India [1] for mea- surements and irradiation tests. The LVRs on the boards were supplied pairwise with V;,, and the load on the output side was chosen such that at V.,,, = 1.8 V the output current was set to around 1.4 A. The Enable pin was set to GND in all cases so that chips were active and could be observed continuously (in Fig. 1 for a pair of ASICs). Figure 1: Thermal image of one of the three PCBs under test with only two LVR chips active. The resistor traces reached more than 60°C without cooling.
Figure 2: Proton beam homogeneity measured with GAFCHROMIC EBT QD+ film and scanned on a densit- ometer. Within the area of LVRs the deposited dose was homogeneous to within + 5%. and connected to a PTW UNIDOS [3] device, which was read out outside of the experimental hall. The LVR de- vices were all irradiated in an active state and their output monitored during all the irradiation. Beam homogeneity (depicted in Fig. 2) was proven by means of Gafchromic EBT QD+ film [4] (blue insert) which was scanned on a densitometer after irradiation. The proton dose varied not more than +5% within an area of 28 mm in diameter. The PCBs were connected to the power supply with a 1.5 mm? thick cable to reduce the voltage drop on the input channel. Additional sense wires were used to monitor the Vj, (of chip pairs) and V,.,, of every chip continuously during the irradiation runs. Every parameter was registered also in a log file, what allows for an analysis of the chips’ behavior before, during and after the exposure.
Figure 2: Proton beam homogeneity measured with GAFCHROMIC EBT QD+ film and scanned on a densit- ometer. Within the area of LVRs the deposited dose was homogeneous to within + 5%. and connected to a PTW UNIDOS [3] device, which was read out outside of the experimental hall. The LVR de- vices were all irradiated in an active state and their output monitored during all the irradiation. Beam homogeneity (depicted in Fig. 2) was proven by means of Gafchromic EBT QD+ film [4] (blue insert) which was scanned on a densitometer after irradiation. The proton dose varied not more than +5% within an area of 28 mm in diameter. The PCBs were connected to the power supply with a 1.5 mm? thick cable to reduce the voltage drop on the input channel. Additional sense wires were used to monitor the Vj, (of chip pairs) and V,.,, of every chip continuously during the irradiation runs. Every parameter was registered also in a log file, what allows for an analysis of the chips’ behavior before, during and after the exposure.
Figure 3: Changes in V,,,; in volt of 8 tested 1.8 V LVR chips with increasing dose (upper) and relative deviation of Vout (normalized to the starting value) in percent (lower part). Values in the upper part have been intentionally shifted by 0.01 V against each other for better readabil- ity. The vertical break-off at late times marks the situation after 25 kGy where some devices were shut down for yet unknown reasons. As four devices shut-down simultane- ously, this event is not interpreted as irradiation induced.
Figure 3: Changes in V,,,; in volt of 8 tested 1.8 V LVR chips with increasing dose (upper) and relative deviation of Vout (normalized to the starting value) in percent (lower part). Values in the upper part have been intentionally shifted by 0.01 V against each other for better readabil- ity. The vertical break-off at late times marks the situation after 25 kGy where some devices were shut down for yet unknown reasons. As four devices shut-down simultane- ously, this event is not interpreted as irradiation induced.
Figure 1: mSTS enclosure with downstream wall removed.
Figure 1: mSTS enclosure with downstream wall removed.
Figure 2: mSTS installed on the beam table in Cave HTD. References
Figure 2: mSTS installed on the beam table in Cave HTD. References
The simulation results for the FEE cooling plate to be used in the thermal demonstrator, i.e. for STS Unit 01, are summarized below and shown in Fig. 2.
The simulation results for the FEE cooling plate to be used in the thermal demonstrator, i.e. for STS Unit 01, are summarized below and shown in Fig. 2.
Figure 1: Placement of nozzles in the C-Frame. STS sensors will be cooled by forced N2 cooling to en- sure that no extra material budget is included inside the ac- tive detector volume. As a preliminary design, gas noz- zles located on the top of individual ladders will be used to cool the respective sensors (see Fig. 1). Cold and dry gas (-30°C; 0.1% RH at 20°C) will be supplied by commercial gas chillers to every nozzle.
Figure 1: Placement of nozzles in the C-Frame. STS sensors will be cooled by forced N2 cooling to en- sure that no extra material budget is included inside the ac- tive detector volume. As a preliminary design, gas noz- zles located on the top of individual ladders will be used to cool the respective sensors (see Fig. 1). Cold and dry gas (-30°C; 0.1% RH at 20°C) will be supplied by commercial gas chillers to every nozzle.
Figure 2: Thermal FEA results for FEE cooling. Note that air convection, radiation, cable heating and thermal interface materials have not been taken into con- sideration for this analysis.
Figure 2: Thermal FEA results for FEE cooling. Note that air convection, radiation, cable heating and thermal interface materials have not been taken into con- sideration for this analysis.
Figure 3: Distribution System Design.
Figure 3: Distribution System Design.
Figure 4: Power cables cooling by FEE box. In order to remove the ~ 2 kW additional heat in the FEE setup, its aimed to derive the needed cooling from existing sources without major geometry changes. So a modified FEE box will be used, where cable cooling is done by a metallic cable cover attached on the corner fin of the FEE box (Fig. 4).
Figure 4: Power cables cooling by FEE box. In order to remove the ~ 2 kW additional heat in the FEE setup, its aimed to derive the needed cooling from existing sources without major geometry changes. So a modified FEE box will be used, where cable cooling is done by a metallic cable cover attached on the corner fin of the FEE box (Fig. 4).
Table 1: CFD analysis results with NOVEC cooling The tube geometry figures are based on the fact that press-fitted tube channels will be used for cooling plate manufacturing. Note that air convection, radiation, cable heating and thermal interface materials have not been taken into consideration for this analysis.
Table 1: CFD analysis results with NOVEC cooling The tube geometry figures are based on the fact that press-fitted tube channels will be used for cooling plate manufacturing. Note that air convection, radiation, cable heating and thermal interface materials have not been taken into consideration for this analysis.
Figure 1: CFD result for 8 FEE boxes dissipating heat on both sides of a cooling plate. The simulation considers a coolant temperature T = -40°C and mass flow 7m = 150 g/s. The color scale shows temperatures in °C.
Figure 1: CFD result for 8 FEE boxes dissipating heat on both sides of a cooling plate. The simulation considers a coolant temperature T = -40°C and mass flow 7m = 150 g/s. The color scale shows temperatures in °C.
Figure 1: Green tracks are reconstructable (can be used for the alignment), where red tracks are not.
Figure 1: Green tracks are reconstructable (can be used for the alignment), where red tracks are not.
Figure 2: Cosmic muon track slope distribution in X direc- tion (T,, green) and in Y direction (J',, red) with respect to the beam axis.
Figure 2: Cosmic muon track slope distribution in X direc- tion (T,, green) and in Y direction (J',, red) with respect to the beam axis.
Figure 5: X directional misalignment correction.
Figure 5: X directional misalignment correction.
Figure 3: Residual distribution in X direction. gorithm to produce the alignment corrections. The align- ment results are discussed below.
Figure 3: Residual distribution in X direction. gorithm to produce the alignment corrections. The align- ment results are discussed below.
Figure 4: Residual distribution in Y direction.
Figure 4: Residual distribution in Y direction.
Figure 6: Y directional misalignment correction.
Figure 6: Y directional misalignment correction.
Figure 1: Online event display from the HADES RICH, no timing cuts applied yet.
Figure 1: Online event display from the HADES RICH, no timing cuts applied yet.
nels of a Trb3Sc with the linear calibration in software (blue) and the calibration on the FPGA (red). The differ- ence in precision between both methods is on a level of 1 ps only verifying the FPGA implementation. Within the errors of the measurements and the implementation, the software and FPGA calibration give comparable results with an out- standing precision. Figure 2: Measured RMS for different channels of a Trb3Sc with a software (blue) and FPGA (red) based linear calibration.
nels of a Trb3Sc with the linear calibration in software (blue) and the calibration on the FPGA (red). The differ- ence in precision between both methods is on a level of 1 ps only verifying the FPGA implementation. Within the errors of the measurements and the implementation, the software and FPGA calibration give comparable results with an out- standing precision. Figure 2: Measured RMS for different channels of a Trb3Sc with a software (blue) and FPGA (red) based linear calibration.
Figure 1: Mean value of the RMS for different measure- ments between 5°C and 50°C of channel 8. The mea- surements were done with a software based bin-by-bin (red), linear (blue) and constant bin-by-bin calibration from 20°C.
Figure 1: Mean value of the RMS for different measure- ments between 5°C and 50°C of channel 8. The mea- surements were done with a software based bin-by-bin (red), linear (blue) and constant bin-by-bin calibration from 20°C.
Figure 2: Loaded mirror supporting frame prototype with indication of the micrometers positions (left). Right: Mi- crometers for measuring horizontal (No. 5) and vertical displacements (No. 7). Next, all components for a full-scale prototype [2] of the mirror supporting frame were manufactured, tested [3] and assembled (Fig. 2). Tests were carried out on the deforma- tion and stability of the structure.
Figure 2: Loaded mirror supporting frame prototype with indication of the micrometers positions (left). Right: Mi- crometers for measuring horizontal (No. 5) and vertical displacements (No. 7). Next, all components for a full-scale prototype [2] of the mirror supporting frame were manufactured, tested [3] and assembled (Fig. 2). Tests were carried out on the deforma- tion and stability of the structure.
Figure 1: Small frame design model (above), photo of the produced frames with imitations of the mirror tiles (bot- tom). A lightweight aluminum full-scale prototype of the mir- ror supporting frame for the CBM RICH detector [1] was produced and tested this year. As a first step, in order to further reduce the amount of materials in detector accep- tance, lightweight small frames and six mounts were pro- duced and tested with mirror tile imitators (Fig. 1).
Figure 1: Small frame design model (above), photo of the produced frames with imitations of the mirror tiles (bot- tom). A lightweight aluminum full-scale prototype of the mir- ror supporting frame for the CBM RICH detector [1] was produced and tested this year. As a first step, in order to further reduce the amount of materials in detector accep- tance, lightweight small frames and six mounts were pro- duced and tested with mirror tile imitators (Fig. 1).
Figure 3: Example of ANSYS deformation calculations un- der load (a), measured displacements during loading (top) and unloading (bottom) of the frame (b), example of dis- placement measurements during three weeks (c). The measurement results are shown in Fig. 3. The re- sults of measurements coincide with calculations within 10-15%. As you can see, during the tests we did not ob- serve any significant displacements of the pillars (maxi- mum 30-40 zm). The test continues.
Figure 3: Example of ANSYS deformation calculations un- der load (a), measured displacements during loading (top) and unloading (bottom) of the frame (b), example of dis- placement measurements during three weeks (c). The measurement results are shown in Fig. 3. The re- sults of measurements coincide with calculations within 10-15%. As you can see, during the tests we did not ob- serve any significant displacements of the pillars (maxi- mum 30-40 zm). The test continues.
Figure 2: Time over threshold for all the channels (operat- ing threshold: 120 mV before (left) and after (right) ToT peak alignment) Figure 1: Time over threshold for all the channels (oper- ating threshold: 30 mV) before (left) and after (right) ToT peak alignment
Figure 2: Time over threshold for all the channels (operat- ing threshold: 120 mV before (left) and after (right) ToT peak alignment) Figure 1: Time over threshold for all the channels (oper- ating threshold: 30 mV) before (left) and after (right) ToT peak alignment
V. Patel, C.Pauly, I. Kres, J. Fértsch, and K-H. Kampert
V. Patel, C.Pauly, I. Kres, J. Fértsch, and K-H. Kampert
Figure 5: Single event plots without ToT cut (left) and after ToT cut (right)(operating threshold: 120 mV) mean is not compromised significantly. This is also con- firmed by single event plots shown in figure 4 and figure 5. In figure 4 we can see a single event plot for threshold of 30mV. Left and right side of Figure 4 show the same event with (right) and without (left) application of the ToT cut. One can easily see that the ring quality is improved by
Figure 5: Single event plots without ToT cut (left) and after ToT cut (right)(operating threshold: 120 mV) mean is not compromised significantly. This is also con- firmed by single event plots shown in figure 4 and figure 5. In figure 4 we can see a single event plot for threshold of 30mV. Left and right side of Figure 4 show the same event with (right) and without (left) application of the ToT cut. One can easily see that the ring quality is improved by
Figure 3: Multiplicity of detected hits without ToT cut (blue) and after applying ToT cut (dotted red) for discrim- ination threshold of 30 mV (left plot) and 120mV (right plot)
Figure 3: Multiplicity of detected hits without ToT cut (blue) and after applying ToT cut (dotted red) for discrim- ination threshold of 30 mV (left plot) and 120mV (right plot)
Figure 4: Single event plots of the same event without ToT cut (left) and after ToT cut (right) (operating threshold: 3 mV)
Figure 4: Single event plots of the same event without ToT cut (left) and after ToT cut (right) (operating threshold: 3 mV)
Figure 2: The first small scaled prototype of the CBM RICH camera with readout electronics. All electronic read- out modules are fixed by using special clamps. On the left side you can see solid copper rails, these are going to be used for the low voltage / high current distribution. Figure 1: Sketch of the latest CBM RICH camera design.
Figure 2: The first small scaled prototype of the CBM RICH camera with readout electronics. All electronic read- out modules are fixed by using special clamps. On the left side you can see solid copper rails, these are going to be used for the low voltage / high current distribution. Figure 1: Sketch of the latest CBM RICH camera design.
under realistic conditions.
under realistic conditions.
The most recent shielding box design (fig. 2) is made from simple plates and weights about 1000 kg. It provides stray field suppression down to below 1 mT which is al- ready close to the design value. Further shape optimization will be performed. Figure 2: Current shielding box design and simulation re- sults. The B, component of the stray field in the lateral plane is suppressed below 1 mT.
The most recent shielding box design (fig. 2) is made from simple plates and weights about 1000 kg. It provides stray field suppression down to below 1 mT which is al- ready close to the design value. Further shape optimization will be performed. Figure 2: Current shielding box design and simulation re- sults. The B, component of the stray field in the lateral plane is suppressed below 1 mT.
Figure 1: “Original” and “layered” field clamp designs and their simulation results.
Figure 1: “Original” and “layered” field clamp designs and their simulation results.
In the simulation the start time of an individual event is generated using the Poisson process. A time stamp of a RICH hit is the sum of the event time, the time of flight from the collision point to the detector, and a measurement error of the detector. For the Cherenkov photons originat- ing from signal e* the time of flight until the photon de- tector plane equals to around 19 ns and is almost the same for all photons. The time resolution of the photon detector was obtained after several lab and in-beam tests and is ex- pected to be less then | ns. In the simulation each RICH hit time is smeared according to a Gaussian distribution with o = 1ns. The dark noise of individual pixels is set to kHz. One can see these noise hits (light blue color) in Figure 1. The event correlated noise depends on the event multiplicity (5 noise hits / MC RICH point / pixel / second). The pixel dead time is set to 50 ns. Figure 1: Distribution of the number of RICH hits in time (UrQMD Au-Au minimum bias, 8 AGeV, 10 MHz). Each color corresponds to an individual event. Light blue color corresponds to the noise hits.
In the simulation the start time of an individual event is generated using the Poisson process. A time stamp of a RICH hit is the sum of the event time, the time of flight from the collision point to the detector, and a measurement error of the detector. For the Cherenkov photons originat- ing from signal e* the time of flight until the photon de- tector plane equals to around 19 ns and is almost the same for all photons. The time resolution of the photon detector was obtained after several lab and in-beam tests and is ex- pected to be less then | ns. In the simulation each RICH hit time is smeared according to a Gaussian distribution with o = 1ns. The dark noise of individual pixels is set to kHz. One can see these noise hits (light blue color) in Figure 1. The event correlated noise depends on the event multiplicity (5 noise hits / MC RICH point / pixel / second). The pixel dead time is set to 50 ns. Figure 1: Distribution of the number of RICH hits in time (UrQMD Au-Au minimum bias, 8 AGeV, 10 MHz). Each color corresponds to an individual event. Light blue color corresponds to the noise hits.
The implemented algorithm was tested in simulation studies performed for UrQMD Au-Au minimum bias col- lisions at 8 AGeV beam energy assuming 10 MHz interac- tion rate (IR). In addition 5 e+ and 5 e~ were embedded into each UrQMD event at the primary vertex. The initial distribution of RICH hits and reconstructed rings in time is shown in Figure 2. The reconstructed rings clearly repre- sent groups, corresponding to events. The ring reconstruc- tion efficiency integrated over the momentum range from 0 — 12 GeV/c is 96.5% for primary e+ and 97.7% for the ref- erence set, i.e. for rings with at least 15 hits (see Figure 3 left). The performance of the ring reconstruction algorithm does not depend on the IR and is the same as the event-by- event reconstruction performance (see Figure 3 right). To summarize, a time-based reconstruction for the RICH de- tector was developed and shows good performance in the simulation studies. Figure 2: Distribution of the number of RICH hits (red) in time and the number of reconstructed rings (blue) in time.
The implemented algorithm was tested in simulation studies performed for UrQMD Au-Au minimum bias col- lisions at 8 AGeV beam energy assuming 10 MHz interac- tion rate (IR). In addition 5 e+ and 5 e~ were embedded into each UrQMD event at the primary vertex. The initial distribution of RICH hits and reconstructed rings in time is shown in Figure 2. The reconstructed rings clearly repre- sent groups, corresponding to events. The ring reconstruc- tion efficiency integrated over the momentum range from 0 — 12 GeV/c is 96.5% for primary e+ and 97.7% for the ref- erence set, i.e. for rings with at least 15 hits (see Figure 3 left). The performance of the ring reconstruction algorithm does not depend on the IR and is the same as the event-by- event reconstruction performance (see Figure 3 right). To summarize, a time-based reconstruction for the RICH de- tector was developed and shows good performance in the simulation studies. Figure 2: Distribution of the number of RICH hits (red) in time and the number of reconstructed rings (blue) in time.
Figure 3: Ring reconstruction efficiency. Left: in depen- dence on momentum. Right: in dependence on IR. ”Zero” IR corresponds to the event-by-event mode.
Figure 3: Ring reconstruction efficiency. Left: in depen- dence on momentum. Right: in dependence on IR. ”Zero” IR corresponds to the event-by-event mode.
References
References
Figure 1: Average gain over all pixels of all measured 1119 MAPM'Ts excluding the reference MAPMT for a high volt- age of —1000V. Orange histogram shows the HAMA- MATSU H8500 64 ch MAPMT for comparison. Bergische Universitat Wuppertal, Germany
Figure 1: Average gain over all pixels of all measured 1119 MAPM'Ts excluding the reference MAPMT for a high volt- age of —1000V. Orange histogram shows the HAMA- MATSU H8500 64 ch MAPMT for comparison. Bergische Universitat Wuppertal, Germany
Figure 1: Left: 4 mirror test set-up after initial alignment and used as reference picture. Right: Lower left mirror ro- tated by 3 mrad around its horizontal axis. The grid appears broken at the edges of the rotated mirror. Mirrors are previously aligned and a reference picture of the mirror wall is taken. On this picture an unbroken grid of retroreflective stripes is observed (Figure la). A rotation is then applied on one mirror tile and another picture of the mirror wall is taken (Figure 1b). On this picture, stripes now appear broken at the edges of the rotated tile. This picture is subsequently subtracted from the reference one. A threshold is applied on the resulting picture to facilitate the measurement of the visible bands of pixels (Figure 2a).
Figure 1: Left: 4 mirror test set-up after initial alignment and used as reference picture. Right: Lower left mirror ro- tated by 3 mrad around its horizontal axis. The grid appears broken at the edges of the rotated mirror. Mirrors are previously aligned and a reference picture of the mirror wall is taken. On this picture an unbroken grid of retroreflective stripes is observed (Figure la). A rotation is then applied on one mirror tile and another picture of the mirror wall is taken (Figure 1b). On this picture, stripes now appear broken at the edges of the rotated tile. This picture is subsequently subtracted from the reference one. A threshold is applied on the resulting picture to facilitate the measurement of the visible bands of pixels (Figure 2a).
Figure 2: Left: Picture obtained after subtracting the ref- erence and rotated pictures and after applying a threshold. Right: Width of the band number 2 for vertical rotations ranging from 0 to 5 mrad (red line) and linear fit (dashed black). Similar results are obtained for all 4 bands. vertical axis, the 4 chosen measurement rows are illustrated on Figure 2a. A linear relation between mirror rotation and band width was found for all mirror tiles of the 4 mirror test set-up. Figure 2b shows the band widths obtained on the second measurement row for several vertical mirror ro- tations.
Figure 2: Left: Picture obtained after subtracting the ref- erence and rotated pictures and after applying a threshold. Right: Width of the band number 2 for vertical rotations ranging from 0 to 5 mrad (red line) and linear fit (dashed black). Similar results are obtained for all 4 bands. vertical axis, the 4 chosen measurement rows are illustrated on Figure 2a. A linear relation between mirror rotation and band width was found for all mirror tiles of the 4 mirror test set-up. Figure 2b shows the band widths obtained on the second measurement row for several vertical mirror ro- tations.
Figure 1: Schematic of the detector setup inside mCBM cave
Figure 1: Schematic of the detector setup inside mCBM cave
Figure 2: Picture of the detector installed inside mCBM cave. Viewing downstream(left) and upstream(right)
Figure 2: Picture of the detector installed inside mCBM cave. Viewing downstream(left) and upstream(right)
Figure 3: Block diagram of data acquisition system
Figure 3: Block diagram of data acquisition system
Figure 4: Top Left: Spill structure plot for all hits from one FEB (in 10 ms bins). Top Right: Variation of duplicate hits time with the channel number. Bottom Left: Spill structure after removing duplicate hits and noisy channels and the inset is for ls bin size. Bottom Right: Distribution of consecutive duplicate hits
Figure 4: Top Left: Spill structure plot for all hits from one FEB (in 10 ms bins). Top Right: Variation of duplicate hits time with the channel number. Bottom Left: Spill structure after removing duplicate hits and noisy channels and the inset is for ls bin size. Bottom Right: Distribution of consecutive duplicate hits
Figure 5: Left: Distribution of pulse height histogram. Right: Consecuitive time difference distribution for one channel (channel 2)
Figure 5: Left: Distribution of pulse height histogram. Right: Consecuitive time difference distribution for one channel (channel 2)
Figure 7: Left: Time difference spectra between one chan- nel to another channel within one FEB. Time difference spectra between GEM and one of the RPC detector (run 48) The event rate for Ag + Au collisions at beam intensity of 10’sec~! has been calculated using equation (1) and it comes out ~7.5 x 10°sec~+. Where Nz is the beam parti- cles/s, N4 is Avagadro number, p is the density, d is thick- ness and A is atomic mass number of the target, and a is the cross-section. CBMROOT simulations with the actual test setup was carried out. The average number of particle per event falling on the GEM1 acceptance is ~4.542. The par- ticle rate (taking into account the event rate and area of the detector) from simulation comes out ~1.793 kHz cm~?.
Figure 7: Left: Time difference spectra between one chan- nel to another channel within one FEB. Time difference spectra between GEM and one of the RPC detector (run 48) The event rate for Ag + Au collisions at beam intensity of 10’sec~! has been calculated using equation (1) and it comes out ~7.5 x 10°sec~+. Where Nz is the beam parti- cles/s, N4 is Avagadro number, p is the density, d is thick- ness and A is atomic mass number of the target, and a is the cross-section. CBMROOT simulations with the actual test setup was carried out. The average number of particle per event falling on the GEM1 acceptance is ~4.542. The par- ticle rate (taking into account the event rate and area of the detector) from simulation comes out ~1.793 kHz cm~?.
Figure 6: Rate (kHz) for each channel of MUCH-XYTER
Figure 6: Rate (kHz) for each channel of MUCH-XYTER
Two such real-size (~80 cm x ~ 40 cm) trapezoidal shaped triple GEM chambers using NS-2(glue-less) tech- nique were fabricated and tested with the Fe®° source at VECC. These modules were installed in the mCBM[3] ex- periment at GSI. Fig.3(a) shows the test setup at VECC. The Fe spectra obtained from this detector is shown in Fig.3(b). Figure 1: Left: Drift PCB and HV layout. Right: Zoomed picture of optocouplers
Two such real-size (~80 cm x ~ 40 cm) trapezoidal shaped triple GEM chambers using NS-2(glue-less) tech- nique were fabricated and tested with the Fe®° source at VECC. These modules were installed in the mCBM[3] ex- periment at GSI. Fig.3(a) shows the test setup at VECC. The Fe spectra obtained from this detector is shown in Fig.3(b). Figure 1: Left: Drift PCB and HV layout. Right: Zoomed picture of optocouplers
Figure 2: a) Picture of large size readout PCB. b) Picture of GEM foil and leakage test. c) Placing GEM foils with a gap of 2mm. d) Stack of GEM foils placed inside the drift PCB. Thanks to CPDA lab of VECC for providing a clean room.
Figure 2: a) Picture of large size readout PCB. b) Picture of GEM foil and leakage test. c) Placing GEM foils with a gap of 2mm. d) Stack of GEM foils placed inside the drift PCB. Thanks to CPDA lab of VECC for providing a clean room.
Figure 3: a) Test setup at VECC lab b) Fe®® spectrum Figure 4: Left: Schematic of test setup for optocoupler Right: Picture of the test setup in lab
Figure 3: a) Test setup at VECC lab b) Fe®® spectrum Figure 4: Left: Schematic of test setup for optocoupler Right: Picture of the test setup in lab
Figure 2: Left: MATLAB measurement of inner radii. Right: measuring outer radii Figure 1: Different areas for which photos were taken (along with their codes)
Figure 2: Left: MATLAB measurement of inner radii. Right: measuring outer radii Figure 1: Different areas for which photos were taken (along with their codes)
GEM foils consists of a regular array of tiny holes of size ~50yum with a pitch of about 140 jm spread accross its en- tire area. Measurement of geometrical parameters such as the hole-diameters and pitch, is necessary for assessing the quality of foil-production. Manual measurements of each such hole thorughout the foil will be extremely time con- suming and just impossible. So as an alternative approach multiple high resolution images of the foil were taken and measurement done using software fits. Fig.1 shows the schematic picture of the real size trapezoidal GEM foil with several image-locations indicated. The images were taken using an optical microscope.
GEM foils consists of a regular array of tiny holes of size ~50yum with a pitch of about 140 jm spread accross its en- tire area. Measurement of geometrical parameters such as the hole-diameters and pitch, is necessary for assessing the quality of foil-production. Manual measurements of each such hole thorughout the foil will be extremely time con- suming and just impossible. So as an alternative approach multiple high resolution images of the foil were taken and measurement done using software fits. Fig.1 shows the schematic picture of the real size trapezoidal GEM foil with several image-locations indicated. The images were taken using an optical microscope.
Figure 3: Top Left: Plot for inner diameter for LL1. Top Right: Plot for outer diameter for LL1 Bottom:Plot for pitch for LL1 tion, these data were accordingly converted to dimensions in ym. A distribution of the inner and outer diameters and pitch measured at a particular position is shown in Fig.3.
Figure 3: Top Left: Plot for inner diameter for LL1. Top Right: Plot for outer diameter for LL1 Bottom:Plot for pitch for LL1 tion, these data were accordingly converted to dimensions in ym. A distribution of the inner and outer diameters and pitch measured at a particular position is shown in Fig.3.
Figure 4: Different areas for which photos were taken (along with their codes)
Figure 4: Different areas for which photos were taken (along with their codes)
Figure 1: Experimental setup nXYTER based self triggered readout electronics has so far, been the main electronics used for testing all our gem detector prototypes for Muon Chambers(MUCH). How- ever, for the final experiment, a specially designed radiation hard ASIC called MUCH-XYTER will be used for read- ing out signals from the first two stations of MUCH. This ASIC has 128 analog front end channels with a dedicated 5- bit ADC per channel for processing the charge pulses from the GEM detector. It can accept both positive and negative charge with configurable e-link and each e-link has a maxi- mum data transmission rate of 320 Mbps. For GEM detec- tors, we use the negative-pulse configuration. In this paper, we report the first test results of a 10 cm X 10 cm triple GEM detector, with a gap configuration of 3-2-2-2 mm, using the first version of MUCH-XYTER(v2.0) front end board(FEB). A ceramic resistive chain was used for power- ing the GEMs. The test setup at VECC is shown in Fig. 1. It consists of the entire DAQ chain including the FEB, AFCK and FLIB (First level interface board). The triple GEM detector was tested with X-rays from >Re sources with Ar/CO2 (70/30) gas mixture. A typical pulse height spectra as obtained is shown in Fig.2. The 32 comparators corresponding to the 32 ADC channels in each of the 128 analog channels were calibrated from 6 fC to 81 fC with a step of 2.5 fC. The peak corresponds to the ma- jor X-ray peak of 5.9 keV from °°Fe source. The spectra is rather wide as compared to those measured with nXYTER, as expected due to coarser ADC resolution. Hence, the ar- gon escape peak is barely visible. Data were taken for vary- ing range GEM voltages to study the gain variation. The detector gain has been obtained by extracting the number of electrons from the measured peak positions in fC(femto Coulombs) and then dividing it by the number of primaries. Fig 3 shows the variation of this detector gain with high
Figure 1: Experimental setup nXYTER based self triggered readout electronics has so far, been the main electronics used for testing all our gem detector prototypes for Muon Chambers(MUCH). How- ever, for the final experiment, a specially designed radiation hard ASIC called MUCH-XYTER will be used for read- ing out signals from the first two stations of MUCH. This ASIC has 128 analog front end channels with a dedicated 5- bit ADC per channel for processing the charge pulses from the GEM detector. It can accept both positive and negative charge with configurable e-link and each e-link has a maxi- mum data transmission rate of 320 Mbps. For GEM detec- tors, we use the negative-pulse configuration. In this paper, we report the first test results of a 10 cm X 10 cm triple GEM detector, with a gap configuration of 3-2-2-2 mm, using the first version of MUCH-XYTER(v2.0) front end board(FEB). A ceramic resistive chain was used for power- ing the GEMs. The test setup at VECC is shown in Fig. 1. It consists of the entire DAQ chain including the FEB, AFCK and FLIB (First level interface board). The triple GEM detector was tested with X-rays from >Re sources with Ar/CO2 (70/30) gas mixture. A typical pulse height spectra as obtained is shown in Fig.2. The 32 comparators corresponding to the 32 ADC channels in each of the 128 analog channels were calibrated from 6 fC to 81 fC with a step of 2.5 fC. The peak corresponds to the ma- jor X-ray peak of 5.9 keV from °°Fe source. The spectra is rather wide as compared to those measured with nXYTER, as expected due to coarser ADC resolution. Hence, the ar- gon escape peak is barely visible. Data were taken for vary- ing range GEM voltages to study the gain variation. The detector gain has been obtained by extracting the number of electrons from the measured peak positions in fC(femto Coulombs) and then dividing it by the number of primaries. Fig 3 shows the variation of this detector gain with high
Figure 2: Fe55 ADC spectra at HV=4450V 'Variable Energy Cyclotron Centre, Kolkata, India; ?Homi Bhabha National Institute, Anushaktinagar, Mumbai, India; 3University of Calcutta, Kolkata, India
Figure 2: Fe55 ADC spectra at HV=4450V 'Variable Energy Cyclotron Centre, Kolkata, India; ?Homi Bhabha National Institute, Anushaktinagar, Mumbai, India; 3University of Calcutta, Kolkata, India
Figure 3: Gain vs HV at calibration step of 1.5 fC
Figure 3: Gain vs HV at calibration step of 1.5 fC
The effect of calibration on detector gain has been stud- ied by calibrating one FEB at three different sets of cal- ibration steps namely, 1.5 fC, 2 fC and 2.5 fC, with the
The effect of calibration on detector gain has been stud- ied by calibrating one FEB at three different sets of cal- ibration steps namely, 1.5 fC, 2 fC and 2.5 fC, with the
Timing information using MUCH XYTER coupled to
Timing information using MUCH XYTER coupled to
threshold charge being set to 6 fC for all cases. For each of the cases, the detector charge corresponding to the peak position of the spectra was measured. It is observed at any particular HV, the gain value for the three di ent calibration remains almost the same, confirming the FEB reads correct charge from the detector irres tive of its calibration step. Fig 3, Fig 4 and Fig 5 s the gain vs HV spectra at 3 different calibration steps. that ffer- that pec- how De- tector response from two different FEBs calibrated at the same bias settings was studied. It was observed that both the FEBs showed almost same pulse height characteristics for the same bias settings and same detector parameters, as evident from Fig 6 and Fig 7. Figure 6: Gain vs HV of FEB no:121
threshold charge being set to 6 fC for all cases. For each of the cases, the detector charge corresponding to the peak position of the spectra was measured. It is observed at any particular HV, the gain value for the three di ent calibration remains almost the same, confirming the FEB reads correct charge from the detector irres tive of its calibration step. Fig 3, Fig 4 and Fig 5 s the gain vs HV spectra at 3 different calibration steps. that ffer- that pec- how De- tector response from two different FEBs calibrated at the same bias settings was studied. It was observed that both the FEBs showed almost same pulse height characteristics for the same bias settings and same detector parameters, as evident from Fig 6 and Fig 7. Figure 6: Gain vs HV of FEB no:121
Figure 5: Gain vs HV at calibration step of 2.5 fC
Figure 5: Gain vs HV at calibration step of 2.5 fC
GEM detector has also been studied. For this, a coinci- dence test setup using two scintillators and a 10 cm x10 cm GEM detector sandwiched between them was carried out. A GB source (°°Sr) was placed on top of one scintillator. The coincidence signals of the two scintillators was put into one channel of the MUCH-XYTER FEB, while signals from detector was acquired by another FEB. Both FEBs were connected to one AFCK, which is supposed to maintain the time synchronization. Fig 8 shows the preliminary results for the time correlation spectra measured at 4700 V. The sigma of this distribution seems to be of the order of 30 ns which is rather high compared to the previous results with nXYTER. The gain dependence study of this time correla- tion spectra is under progress.
GEM detector has also been studied. For this, a coinci- dence test setup using two scintillators and a 10 cm x10 cm GEM detector sandwiched between them was carried out. A GB source (°°Sr) was placed on top of one scintillator. The coincidence signals of the two scintillators was put into one channel of the MUCH-XYTER FEB, while signals from detector was acquired by another FEB. Both FEBs were connected to one AFCK, which is supposed to maintain the time synchronization. Fig 8 shows the preliminary results for the time correlation spectra measured at 4700 V. The sigma of this distribution seems to be of the order of 30 ns which is rather high compared to the previous results with nXYTER. The gain dependence study of this time correla- tion spectra is under progress.
Figure 9: ADC spectra wihin 400 ns time window Figure 8: Time Correlation spectra On some occasions, two peaks were observed in the time correlation spectra. This is yet to be understood. One rea- son could be that the time synchronization gets disturbed during the course of data taking. Further investigation is on to address this issue. Fig 9 shows the pulse height spectra from all the channels combined for a time window of 400 ns after each scintillator trigger.
Figure 9: ADC spectra wihin 400 ns time window Figure 8: Time Correlation spectra On some occasions, two peaks were observed in the time correlation spectra. This is yet to be understood. One rea- son could be that the time synchronization gets disturbed during the course of data taking. Further investigation is on to address this issue. Fig 9 shows the pulse height spectra from all the channels combined for a time window of 400 ns after each scintillator trigger.
Identical method has been followed to analyze the cur- rent signals induced on the RPC readout strip, as ob- tained from simulation and experiment to find the produced charge and detector efficiency. Each current signal was in- tegrated within a 50 ns window to find the charge induced on the readout. A typical charge distribution obtained for 1000 events from simulation is shown in Fig. 1. To have Figure 1: A typical charge distribution obtained from sim- ulation and scheme of calculating the detector efficiency.
Identical method has been followed to analyze the cur- rent signals induced on the RPC readout strip, as ob- tained from simulation and experiment to find the produced charge and detector efficiency. Each current signal was in- tegrated within a 50 ns window to find the charge induced on the readout. A typical charge distribution obtained for 1000 events from simulation is shown in Fig. 1. To have Figure 1: A typical charge distribution obtained from sim- ulation and scheme of calculating the detector efficiency.
The variation of mean charge produced within RPC gas gap with the applied field, as found from simulation is shown using a black dashed line in Fig. 2 for the gas mix- Figure 2: Simulated result on variation of RPC efficiency (solid lines) with the applied field with gas mixture contain- ing 0.3% SFg. The black dashed line shows the variation of mean charge.
The variation of mean charge produced within RPC gas gap with the applied field, as found from simulation is shown using a black dashed line in Fig. 2 for the gas mix- Figure 2: Simulated result on variation of RPC efficiency (solid lines) with the applied field with gas mixture contain- ing 0.3% SFg. The black dashed line shows the variation of mean charge.
ture containing 0.3% SFg along with 4.5% i-C4Hio and 95.2% C2H2F4. To get the mean produced charge less than (Qmax=) 100 fC the detector needs to be operated at a field of 39.58 kVcm—! (HV~7.9 kV). The variation of detec- tor efficiency in crossing different charge thresholds with the applied field is also shown in Fig. 2. Simulation as- sumes that the noise is completely suppressed and is an ideal case. In our experiment, value of Qj, =30 fC or higher depending on the noise level. Initial experimental Figure 3: Experimental result on variation of RPC effi- ciency (solid lines) with the applied voltage for the gas mix- ture containing 0.3% SF. The black dashed line shows the variation of mean charge.
ture containing 0.3% SFg along with 4.5% i-C4Hio and 95.2% C2H2F4. To get the mean produced charge less than (Qmax=) 100 fC the detector needs to be operated at a field of 39.58 kVcm—! (HV~7.9 kV). The variation of detec- tor efficiency in crossing different charge thresholds with the applied field is also shown in Fig. 2. Simulation as- sumes that the noise is completely suppressed and is an ideal case. In our experiment, value of Qj, =30 fC or higher depending on the noise level. Initial experimental Figure 3: Experimental result on variation of RPC effi- ciency (solid lines) with the applied voltage for the gas mix- ture containing 0.3% SF. The black dashed line shows the variation of mean charge.
Figure 2: The signal wires tension distribution. Figure 1: The general view of the rectangular straw cham- ber prototype. A small rectangular chamber prototype with the working area of 500x500 mm has been produced at LHEP JINR. It consists of two layers, each of them contains 80 straws. The diameter of the straws is 6 mm. The anodes of the chamber is a gilded tungsten wire with a diameter of 30 jum. The thickness of the straw walls is about 60 4m. The
Figure 2: The signal wires tension distribution. Figure 1: The general view of the rectangular straw cham- ber prototype. A small rectangular chamber prototype with the working area of 500x500 mm has been produced at LHEP JINR. It consists of two layers, each of them contains 80 straws. The diameter of the straws is 6 mm. The anodes of the chamber is a gilded tungsten wire with a diameter of 30 jum. The thickness of the straw walls is about 60 4m. The
Straw tube detectors are considered as the option for the 3-rd and 4-th stations of CBM Muon Chamber [1]. The full size octagonal straw module prototype has been produced and tested at LHEP JINR [2]. The straw detector with the rectangular shape [3] can be also used at CBM to reduce the cost and to improve the quality of the muon tracker.
Straw tube detectors are considered as the option for the 3-rd and 4-th stations of CBM Muon Chamber [1]. The full size octagonal straw module prototype has been produced and tested at LHEP JINR [2]. The straw detector with the rectangular shape [3] can be also used at CBM to reduce the cost and to improve the quality of the muon tracker.
High-voltage measurements were carried out at 2000 V with the air. The value of the current is 5 nA. The gas vol- umes of all tubes of the chamber and gas manifold are 9.5 1 and 2 1, respectively. The test of the gas leakage was carried out in the air. The leaks were less than 10 cm?/min. The results of the measurement of the signal wires tension is shown in Fig.2. The average value of the tension is ~80 g. Figure 3: The dependence of the signal amplitude on the high voltage.
High-voltage measurements were carried out at 2000 V with the air. The value of the current is 5 nA. The gas vol- umes of all tubes of the chamber and gas manifold are 9.5 1 and 2 1, respectively. The test of the gas leakage was carried out in the air. The leaks were less than 10 cm?/min. The results of the measurement of the signal wires tension is shown in Fig.2. The average value of the tension is ~80 g. Figure 3: The dependence of the signal amplitude on the high voltage.
Figure 1: Front-end electronics PCB. Front-end electronics board for straw tube detector with 32- charge- sensitive amplifiers based on AST-1-1 chip [2] is shown in Fig.1. These chips are specially designed to read the signals from straw detectors. Each AST-1-1 chip processes signals from eight straw tubes providing eight LVDS outputs to the TDC.
Figure 1: Front-end electronics PCB. Front-end electronics board for straw tube detector with 32- charge- sensitive amplifiers based on AST-1-1 chip [2] is shown in Fig.1. These chips are specially designed to read the signals from straw detectors. Each AST-1-1 chip processes signals from eight straw tubes providing eight LVDS outputs to the TDC.
Figure 2: AST-1-1 ASIC block diagrams. The main features of the AST-1-1 ASIC are eight sig- nal channels, control of input resistance of preamplifier,
Figure 2: AST-1-1 ASIC block diagrams. The main features of the AST-1-1 ASIC are eight sig- nal channels, control of input resistance of preamplifier,
Figure 4: Peaking time of the BLR vs detector capacitance. Figure 3: Differential gain of the BLR vs detector capaci- tance. The dependencies of the differential gain and the peak- ing time of the BLR on the detector capacitance at a supply voltage of 3.3 V are shown in Fig.3 and Fig.4, respectively. The measurements were performed with the default pream-
Figure 4: Peaking time of the BLR vs detector capacitance. Figure 3: Differential gain of the BLR vs detector capaci- tance. The dependencies of the differential gain and the peak- ing time of the BLR on the detector capacitance at a supply voltage of 3.3 V are shown in Fig.3 and Fig.4, respectively. The measurements were performed with the default pream-
adjustment of the ion signal tail, BLR, two modes of dis- criminator operation: time over threshold (ToT) and con- trol of the output signal width, LVDS driver, presence of the ninth channel for monitoring and signal shaping from the straw. The block diagram of the signal channel is given in Fig.2. The preamplifiers has charge sensitive configura- tion. Preamplifiers, shaper, BLR are bipolar circuit, while discriminator, monostable and LVDS drivers have CMOS schematic solution. The minimum operating voltage for the AST-1-1 ASIC is 2.5 V. The optimum combination of parameters and power consumption is achieved at a supply voltage of 2.8 V.
adjustment of the ion signal tail, BLR, two modes of dis- criminator operation: time over threshold (ToT) and con- trol of the output signal width, LVDS driver, presence of the ninth channel for monitoring and signal shaping from the straw. The block diagram of the signal channel is given in Fig.2. The preamplifiers has charge sensitive configura- tion. Preamplifiers, shaper, BLR are bipolar circuit, while discriminator, monostable and LVDS drivers have CMOS schematic solution. The minimum operating voltage for the AST-1-1 ASIC is 2.5 V. The optimum combination of parameters and power consumption is achieved at a supply voltage of 2.8 V.
The matching of the straw tubes impedance can be done by the changing preamplifier input resistance by the solder- ing of resistors on the board. The compensation for the ion tail of the straw signals can be provided by the soldering of the resistors on the amplifier board. The nominal of the re- sistors depends on the design features of the straw chamber and the used gas mixture and can be defined when testing straw chambers. Figure 6: Output LVDS signal width for the ToT option vs the anode voltage.
The matching of the straw tubes impedance can be done by the changing preamplifier input resistance by the solder- ing of resistors on the board. The compensation for the ion tail of the straw signals can be provided by the soldering of the resistors on the amplifier board. The nominal of the re- sistors depends on the design features of the straw chamber and the used gas mixture and can be defined when testing straw chambers. Figure 6: Output LVDS signal width for the ToT option vs the anode voltage.
plifier input impedance of R;,= 120 Ohm and Q;,,=10 fC. The values of the differential gain and the peaking time at Cy=15 pF are ~22 and ~6 ns, respectively.
plifier input impedance of R;,= 120 Ohm and Q;,,=10 fC. The values of the differential gain and the peaking time at Cy=15 pF are ~22 and ~6 ns, respectively.
is ~100 ns at 1500 V. Another option is the fixed width of the output LVDS signal. The rate capability of the FEE has been checked using a generator signal with a duration of 20 ns and an amplitude of 10 mV fed to 32 channels of the board through 1 kOhm. The maximal rate allowed is ~10 MHz for the width of the LVDS signal width of 60 ns and a threshold setting of 100 mV. Figure 7: The counts profile from the straw chamber FEE.
is ~100 ns at 1500 V. Another option is the fixed width of the output LVDS signal. The rate capability of the FEE has been checked using a generator signal with a duration of 20 ns and an amplitude of 10 mV fed to 32 channels of the board through 1 kOhm. The maximal rate allowed is ~10 MHz for the width of the LVDS signal width of 60 ns and a threshold setting of 100 mV. Figure 7: The counts profile from the straw chamber FEE.
Figure 8: Drift time for 32 straws.
Figure 8: Drift time for 32 straws.
Figure 1: Measured Efficiency with respect to high voltage using commercial CSA 142]H. The prototype bakelite RPC has shown > 95% efficiency in detecting cosmic muons at 10.5 kV [1] using the follow-
Figure 1: Measured Efficiency with respect to high voltage using commercial CSA 142]H. The prototype bakelite RPC has shown > 95% efficiency in detecting cosmic muons at 10.5 kV [1] using the follow-
kV. The charge sepctra of the detetctor has also been mea- sured at 9.8 kV using MANAS based electronics [2] and the maximum measured charge was ~ 270 fC as shown in Fig.2. Figure 2: Measured Charge Spectrum using MANAS DAQ.
kV. The charge sepctra of the detetctor has also been mea- sured at 9.8 kV using MANAS based electronics [2] and the maximum measured charge was ~ 270 fC as shown in Fig.2. Figure 2: Measured Charge Spectrum using MANAS DAQ.
137 Figure 3: Detetctors setup and position of Cs~°* source in GIF++ 4.5 : 0.3. The experimental setup is shown in Fig.3. The RPC was placed 1.67 m downstream the Cs!8" source be- hind two CBM-TRD detetctors. A finger scintillator (3 cm x 12 cm) has been placed in front of the TRD detectors so that it’s projection completely overlapped with the proto- type RPC detector. Two big size (30 cm x 40 cm) scintil-
137 Figure 3: Detetctors setup and position of Cs~°* source in GIF++ 4.5 : 0.3. The experimental setup is shown in Fig.3. The RPC was placed 1.67 m downstream the Cs!8" source be- hind two CBM-TRD detetctors. A finger scintillator (3 cm x 12 cm) has been placed in front of the TRD detectors so that it’s projection completely overlapped with the proto- type RPC detector. Two big size (30 cm x 40 cm) scintil-
Figure 4: Efficiency of muon detetction of the prototype RPC in source OFF consition as a function of high voltage in GIF++. As shown in Fig.4, the detetctor shows a muon efficiency > 95% at 9.1 kV in source OFF condition. We have also
Figure 4: Efficiency of muon detetction of the prototype RPC in source OFF consition as a function of high voltage in GIF++. As shown in Fig.4, the detetctor shows a muon efficiency > 95% at 9.1 kV in source OFF condition. We have also
Figure 5: Hit Rate of muon beam on the whole RPC detec- tor at 9200V in source OFF condition.
Figure 5: Hit Rate of muon beam on the whole RPC detec- tor at 9200V in source OFF condition.
Figure 6: Efficiency of the detetctor for detecting muons in highly irradiated gamma background. Fig.6 shows the efficiency with high voltage in pres- ence of gamma background. The photon flux falling on the detector has been calculated for diffferent attenuation fac- tors. The efficiency is seen to be reduced in presence of gamma. At 9.6 kV the detector has shown > 90 % muon efficiency when the photon flux is 40 kHz/cm?. Further analysis of the test beam results is ongoing.
Figure 6: Efficiency of the detetctor for detecting muons in highly irradiated gamma background. Fig.6 shows the efficiency with high voltage in pres- ence of gamma background. The photon flux falling on the detector has been calculated for diffferent attenuation fac- tors. The efficiency is seen to be reduced in presence of gamma. At 9.6 kV the detector has shown > 90 % muon efficiency when the photon flux is 40 kHz/cm?. Further analysis of the test beam results is ongoing.
Figure 1: Variation of gain, energy resolution and T/p as a function of time obtain after the Gaussian fitting of the spectra. The spec- tra are stored automatically using ORTEC MCA at a regu- lar interval of 10 minutes and CuteCom software package has been used for continuous monitoring of the tempera- ture and pressure around the chamber. The measurement of gain and energy resolution has been carried out uninter- ruptedly for a period of >1200 hours. Figure 1 Shows the variation of measured gain, energy resolution and £ > asa function of time.
Figure 1: Variation of gain, energy resolution and T/p as a function of time obtain after the Gaussian fitting of the spectra. The spec- tra are stored automatically using ORTEC MCA at a regu- lar interval of 10 minutes and CuteCom software package has been used for continuous monitoring of the tempera- ture and pressure around the chamber. The measurement of gain and energy resolution has been carried out uninter- ruptedly for a period of >1200 hours. Figure 1 Shows the variation of measured gain, energy resolution and £ > asa function of time.
Figure 2: Gain (top left), count rate (top right), energy res- olution (bottom), at 20 different places on the detector Since it is not possible to scale all the measured parame- ters for a large area triple GEM detector, it is in our future plan to fabricate a large area triple GEM detector and to perform the stability and basic characteristics study of the detector.
Figure 2: Gain (top left), count rate (top right), energy res- olution (bottom), at 20 different places on the detector Since it is not possible to scale all the measured parame- ters for a large area triple GEM detector, it is in our future plan to fabricate a large area triple GEM detector and to perform the stability and basic characteristics study of the detector.
Figure 1: Noise rate and Efficiency Vs. voltage The possibility of using single gap Resistive Plate Cham- ber (RPC) in the 3"@ and 4*” stations of the CBM-MUCH detector are currently being explored. RPC detector tech- nology is widely used in high energy physics experi- ments, for trigger and tracking purposes, due to its ex- cellent efficiency (> 90%) and time resolution (1-2 ns). The maximum particle flux on the 3"¢ and 4*” stations of CBM-MUCH have been estimated to be 15 kHz/cm? and 5.6 kHz/cm?, respectively, for minimum bias Au—Au colli- sions at 8 AGeV. Therefore, the inexpensive RPC’s, known for their moderate rate handling capacity (~ 1-2 kHz/cm?, depending on the resistivity of the material used), are be- ing considered for the CBM. We are currently investigat- ing the possibility of using low-resistive materials for RPC fabrication that will ensure stable operation at higher rates (~ 10 kHz/cm?) in the avalanche mode.
Figure 1: Noise rate and Efficiency Vs. voltage The possibility of using single gap Resistive Plate Cham- ber (RPC) in the 3"@ and 4*” stations of the CBM-MUCH detector are currently being explored. RPC detector tech- nology is widely used in high energy physics experi- ments, for trigger and tracking purposes, due to its ex- cellent efficiency (> 90%) and time resolution (1-2 ns). The maximum particle flux on the 3"¢ and 4*” stations of CBM-MUCH have been estimated to be 15 kHz/cm? and 5.6 kHz/cm?, respectively, for minimum bias Au—Au colli- sions at 8 AGeV. Therefore, the inexpensive RPC’s, known for their moderate rate handling capacity (~ 1-2 kHz/cm?, depending on the resistivity of the material used), are be- ing considered for the CBM. We are currently investigat- ing the possibility of using low-resistive materials for RPC fabrication that will ensure stable operation at higher rates (~ 10 kHz/cm?) in the avalanche mode.
Figure 1: Variation of gain and T/p as a function of time
Figure 1: Variation of gain and T/p as a function of time
Figure 2: Variation of normalized gain as a function of ac- cumulated charge
Figure 2: Variation of normalized gain as a function of ac- cumulated charge
Figure 1: Test Bench Setup for Calibration at VECC The new dual gain STS/MUCH-XYTER 1s a 128 chan- nel highly configurable ASIC with dedicated 5-bit flash ADC for each individual channel. This ADC have 31 com- parators which can be trimmed for a particular reference voltage controlled by a 8-bit Digital to Analog Convertor (DAC). Each bit needs to be calibrated for a particular input charge which can be either linear or non-linear within its dynamic range. A known charge is injected for any given ADC-channel and then trimming offset of that particular channel is set such that ADC decoder output shows the tar- geted channel only. A counter feedback algorithm is also developed to digitally adjust the comparator offset to mini- mize the deviation from the expected value. To calibrate 31 comparators of an ADC of all 128 channels in the MUCH mode, it was required to inject the known charge at the in- put of the channel. This charge was injected using a voltage to charge convertor circuit controlled by a pulse generator. If manually done, this is a time taking process which re- quired at least 31 time interventions to complete the cali- bration. Therefore we have developed an automated com- parator calibration technique which includes automation of several equipment used in this setup.
Figure 1: Test Bench Setup for Calibration at VECC The new dual gain STS/MUCH-XYTER 1s a 128 chan- nel highly configurable ASIC with dedicated 5-bit flash ADC for each individual channel. This ADC have 31 com- parators which can be trimmed for a particular reference voltage controlled by a 8-bit Digital to Analog Convertor (DAC). Each bit needs to be calibrated for a particular input charge which can be either linear or non-linear within its dynamic range. A known charge is injected for any given ADC-channel and then trimming offset of that particular channel is set such that ADC decoder output shows the tar- geted channel only. A counter feedback algorithm is also developed to digitally adjust the comparator offset to mini- mize the deviation from the expected value. To calibrate 31 comparators of an ADC of all 128 channels in the MUCH mode, it was required to inject the known charge at the in- put of the channel. This charge was injected using a voltage to charge convertor circuit controlled by a pulse generator. If manually done, this is a time taking process which re- quired at least 31 time interventions to complete the cali- bration. Therefore we have developed an automated com- parator calibration technique which includes automation of several equipment used in this setup.
dividual channel. In Figure 1, the test bench setup picture is shown and in Figure 2, the picture of MUCH-XYTER FEE board and compact connector with charge injection circuit for individual channel is also there at top of the im- age. Other components includes a Kintex-7 FPGA based data processing board. Back-end of this board has an IPbus interface via which we control and readback all the trim- ming commands as well as data. After a defined charge is injected to all the channels, data is readback from the out- put buffer register of all the individual ADCs. A counter feedback mechanism is then used to digitally adjust the comparator’s offset which is controlled by a trimming reg- ister so that the offset is minimized from the desired value. Apart from the ADC calibration, we have also developed a Voltage Scanning procedure to determine proper bias set- tings to calibrate MUCH-XYTER ASIC. Figure 2: FEE Board with MUCH-XYTER ASIC and con- nector for charge injection
dividual channel. In Figure 1, the test bench setup picture is shown and in Figure 2, the picture of MUCH-XYTER FEE board and compact connector with charge injection circuit for individual channel is also there at top of the im- age. Other components includes a Kintex-7 FPGA based data processing board. Back-end of this board has an IPbus interface via which we control and readback all the trim- ming commands as well as data. After a defined charge is injected to all the channels, data is readback from the out- put buffer register of all the individual ADCs. A counter feedback mechanism is then used to digitally adjust the comparator’s offset which is controlled by a trimming reg- ister so that the offset is minimized from the desired value. Apart from the ADC calibration, we have also developed a Voltage Scanning procedure to determine proper bias set- tings to calibrate MUCH-XYTER ASIC. Figure 2: FEE Board with MUCH-XYTER ASIC and con- nector for charge injection
Figure 3: Online Power Monitoring System for Agilent Power Supply There are several occasion when the power cycle is re- quired for some particular channel of LV. Apart from lab testing, during the beam time, these low voltage power supplies are generally housed close to the detector which are not accessible during the beam time and required to be accessed frequently for a clean start of the readout chain. Hence, an automated Control System and Constant Power Monitoring System has been developed by which we can control power and current of the power supply. We can even monitor the status of currents and voltages of all in- dividual channels of the power supply. Same as the Tek- tronix function generator, python package called Py-VISA and backend library called Py-VISA-py has been used to develop this Control and Monitoring setup. A Python script has been written using module specific GPIB commands to access this Agilent power supply for its each channel. A separate text file exists where the list of voltage and cur- rent values are written in accordance with each channel and the power system channels are controlled using the these scripts by fetching the values of voltage and current from this file. The values in the text file can be changed as per the requirement. A screen shot of sample power monitor- ing system is given in Figure 3.
Figure 3: Online Power Monitoring System for Agilent Power Supply There are several occasion when the power cycle is re- quired for some particular channel of LV. Apart from lab testing, during the beam time, these low voltage power supplies are generally housed close to the detector which are not accessible during the beam time and required to be accessed frequently for a clean start of the readout chain. Hence, an automated Control System and Constant Power Monitoring System has been developed by which we can control power and current of the power supply. We can even monitor the status of currents and voltages of all in- dividual channels of the power supply. Same as the Tek- tronix function generator, python package called Py-VISA and backend library called Py-VISA-py has been used to develop this Control and Monitoring setup. A Python script has been written using module specific GPIB commands to access this Agilent power supply for its each channel. A separate text file exists where the list of voltage and cur- rent values are written in accordance with each channel and the power system channels are controlled using the these scripts by fetching the values of voltage and current from this file. The values in the text file can be changed as per the requirement. A screen shot of sample power monitor- ing system is given in Figure 3.
size should remain same as per the calibration file parame- ter (e.g. 2 fC in general) which is governed by AFG 3252 function generator and charge injector circuit on the mating connectors of the FEE boards as shown in Figure 2. After injecting these charges periodically for a pre defined time (e.g. 1 second), all the counter values associated with 31 comparators of that channel are read back via IPBus pro- tocol. Then an ADC linearity plot is generated using the Figure 4: ADC linearity output after voltage scan
size should remain same as per the calibration file parame- ter (e.g. 2 fC in general) which is governed by AFG 3252 function generator and charge injector circuit on the mating connectors of the FEE boards as shown in Figure 2. After injecting these charges periodically for a pre defined time (e.g. 1 second), all the counter values associated with 31 comparators of that channel are read back via IPBus pro- tocol. Then an ADC linearity plot is generated using the Figure 4: ADC linearity output after voltage scan
The LVDB board with its components is shown in figure 1. One of its Channel will be used for optocoupler con- trol rest of all for powering FEE boards of GEM detector. Pressure sensor, humidity sensor and temperature sensor are mounted to detect any leak in the detector. Figure 2: Simplified Control Scheme
The LVDB board with its components is shown in figure 1. One of its Channel will be used for optocoupler con- trol rest of all for powering FEE boards of GEM detector. Pressure sensor, humidity sensor and temperature sensor are mounted to detect any leak in the detector. Figure 2: Simplified Control Scheme
Figure 1: Mass distribution of wrongly identified 4 He for 10’ simulated UrQMD-events with embedded 3H and d. For the identification of ,%He, which decays as ,$He —+ >He + p+ and subsequently as He > *He + p + 7, the separation of d and “He is particularly important. In fact, without clearly separating these two nuclei, it is not possible to distinguish the ° He-decay from the decay of the hypertriton: {He + d+p+77~. As the expected production rate of 3H (1.7 -10? per central Au-Au colli- sion at 8 AGeV [3, 4]) is significantly higher compared to the production rate of ,¢He (4.8-10~® per event), a high background is expected. Figure 1 shows the distribution of wrongly identified He (in blue) in 10” simulated UrQMD- events, each containing one 3H and six d [4], but neither aHe nor ,¢He. With a TOF mass-cut Am= +3 o around the 5 He-mass (in violet) the number of wrongly identified ‘He is around 1,000. The TRD provides a suppression of nearly all misidentified 5 He (in orange). At the same time, the efficiency « = Teh [estore i.e. the ratio of correctly identified particles nM@° sig. to simulated
Figure 1: Mass distribution of wrongly identified 4 He for 10’ simulated UrQMD-events with embedded 3H and d. For the identification of ,%He, which decays as ,$He —+ >He + p+ and subsequently as He > *He + p + 7, the separation of d and “He is particularly important. In fact, without clearly separating these two nuclei, it is not possible to distinguish the ° He-decay from the decay of the hypertriton: {He + d+p+77~. As the expected production rate of 3H (1.7 -10? per central Au-Au colli- sion at 8 AGeV [3, 4]) is significantly higher compared to the production rate of ,¢He (4.8-10~® per event), a high background is expected. Figure 1 shows the distribution of wrongly identified He (in blue) in 10” simulated UrQMD- events, each containing one 3H and six d [4], but neither aHe nor ,¢He. With a TOF mass-cut Am= +3 o around the 5 He-mass (in violet) the number of wrongly identified ‘He is around 1,000. The TRD provides a suppression of nearly all misidentified 5 He (in orange). At the same time, the efficiency « = Teh [estore i.e. the ratio of correctly identified particles nM@° sig. to simulated
particles Nsim, is hardly effected by a (dEdx)-cut. Fig. 2 shows the mass distribution of ,{He for 10° simulated A\He-events embedded in UrQMD-events with d. The re- constructed signal (in blue) is compared to the cases with mass-cut (in violet), the reconstructed particles with correct MC-Id (in red) and with TRD-cut (in orange). The suppres- sion of correctly identified , § He via a 2o-(d dz) -cut is on a negligible level. Figure 2: Mass distribution of ,°He for 10° simulated UrQMD-events with embedded ASH and d.
particles Nsim, is hardly effected by a (dEdx)-cut. Fig. 2 shows the mass distribution of ,{He for 10° simulated A\He-events embedded in UrQMD-events with d. The re- constructed signal (in blue) is compared to the cases with mass-cut (in violet), the reconstructed particles with correct MC-Id (in red) and with TRD-cut (in orange). The suppres- sion of correctly identified , § He via a 2o-(d dz) -cut is on a negligible level. Figure 2: Mass distribution of ,°He for 10° simulated UrQMD-events with embedded ASH and d.
Figure 1: ADC spectra of the samples used for the three proposed methods for baseline determination, each fitted with a Gaussian function (Xe/CO2(80:20), Anode HV = 2000 V). Institut fiir Kernphysik, Miinster, Germany
Figure 1: ADC spectra of the samples used for the three proposed methods for baseline determination, each fitted with a Gaussian function (Xe/CO2(80:20), Anode HV = 2000 V). Institut fiir Kernphysik, Miinster, Germany
Figure 2: Maximum ADC value plotted against the ADC value of sample[31] for self or neighbour triggered mes- sages on all channels (Xe/CO2(80:20), Anode HV = 2000 V).
Figure 2: Maximum ADC value plotted against the ADC value of sample[31] for self or neighbour triggered mes- sages on all channels (Xe/CO2(80:20), Anode HV = 2000 V).
Figure 3: Baseline positions (upper panel) and widths, determined as 20 of the Gaussian fit (lower panel) of channel 6 determined by Gaussian fits for the sample[0], sample[1] and sample[31] methods plotted against time (Xe/CO2(80:20), Anode HV = 2000 V). beam for half of the time recored in these files. As a TSA file is limited in data size and not in time of recording, these three files span over longer time than the other files, as indi- cated by the larger horizontal error bars. Why the baseline position and especially its width drops significantly at a PE- TRA III top-up is not yet fully clear. While it could also be a reconstruction artefact, it would otherwise suggest a load dependence, though a verification of this is non-trivial. If a higher load, quantified by an increased hit rate, would raise baseline position and width, so would a raised base- line position and width increase the hit rate, as absolute thresholds were used. For a disentanglement of these two effects, more refined measurements and analysis have to be conducted. A test setup, specifically for in depth baseline measurements, is currently being installed in the laboratory in Miinster.
Figure 3: Baseline positions (upper panel) and widths, determined as 20 of the Gaussian fit (lower panel) of channel 6 determined by Gaussian fits for the sample[0], sample[1] and sample[31] methods plotted against time (Xe/CO2(80:20), Anode HV = 2000 V). beam for half of the time recored in these files. As a TSA file is limited in data size and not in time of recording, these three files span over longer time than the other files, as indi- cated by the larger horizontal error bars. Why the baseline position and especially its width drops significantly at a PE- TRA III top-up is not yet fully clear. While it could also be a reconstruction artefact, it would otherwise suggest a load dependence, though a verification of this is non-trivial. If a higher load, quantified by an increased hit rate, would raise baseline position and width, so would a raised base- line position and width increase the hit rate, as absolute thresholds were used. For a disentanglement of these two effects, more refined measurements and analysis have to be conducted. A test setup, specifically for in depth baseline measurements, is currently being installed in the laboratory in Miinster.
Figure 4: Baseline positions (upper panel) and widths, de- termined as 2c of the Gaussian fit, (lower panel) of the channels 4 to 7, calculated with the sample[0] method plotted against time. The channels 4 and 6 were lo- cated in the beam spot, the channels 5 and 7 outside of it (Xe/CO2(80:20), Anode HV = 2000 V). have very similarly to each other, one can conclude that the baselines of neighbouring pads are correlated. As chan- nel 4 and 6 were located in the beam spot, thus receiving higher statistics, they have generally less statistical fluctu- ations, but are affected more strongly by the PETRA top- ups. Generally, the width of the baseline needs to be re- duced substantially to achieve the design value. We note that a good detector energy resolution was measured in the lab with a °°Fe source [3].
Figure 4: Baseline positions (upper panel) and widths, de- termined as 2c of the Gaussian fit, (lower panel) of the channels 4 to 7, calculated with the sample[0] method plotted against time. The channels 4 and 6 were lo- cated in the beam spot, the channels 5 and 7 outside of it (Xe/CO2(80:20), Anode HV = 2000 V). have very similarly to each other, one can conclude that the baselines of neighbouring pads are correlated. As chan- nel 4 and 6 were located in the beam spot, thus receiving higher statistics, they have generally less statistical fluctu- ations, but are affected more strongly by the PETRA top- ups. Generally, the width of the baseline needs to be re- duced substantially to achieve the design value. We note that a good detector energy resolution was measured in the lab with a °°Fe source [3].
The most inner 10 TRD modules of each layer of the TRD wall in the CBM experiment are exposed to of particles often in pile-up regime. To disentangle individ- ual hits and reconstruct incident position and energy de- posit, the Bucharest-prototype is designed around two fea- tures: triangular-shaped pad-plane for 2D position sensitiv- ity and non-diagonal self-triggered singled-value read-out ASIC - FASP (Fast Analog Signal Processor) [1 mal data load. An almost exact replica of the C ules wil be installed in the mCBM setup [2] at S high rates for opti- BM mod- S18. The prototype of 59 x 59 cm? with 2880 readout channels is operated For the by 180 FASPs mounted on 30 FASPRO boards. mCBM experiment a new FASP version (v03) is prepared and a 6-FASPs FASPRO (FASP Read-Out) board has been produced and tested (v02). Figure 1: The third version of the FASP ASIC prepared for mCBM test in the design phase (left) and after bonding and ready for integration (right).
The most inner 10 TRD modules of each layer of the TRD wall in the CBM experiment are exposed to of particles often in pile-up regime. To disentangle individ- ual hits and reconstruct incident position and energy de- posit, the Bucharest-prototype is designed around two fea- tures: triangular-shaped pad-plane for 2D position sensitiv- ity and non-diagonal self-triggered singled-value read-out ASIC - FASP (Fast Analog Signal Processor) [1 mal data load. An almost exact replica of the C ules wil be installed in the mCBM setup [2] at S high rates for opti- BM mod- S18. The prototype of 59 x 59 cm? with 2880 readout channels is operated For the by 180 FASPs mounted on 30 FASPRO boards. mCBM experiment a new FASP version (v03) is prepared and a 6-FASPs FASPRO (FASP Read-Out) board has been produced and tested (v02). Figure 1: The third version of the FASP ASIC prepared for mCBM test in the design phase (left) and after bonding and ready for integration (right).
cessing of the chip are transparent from the curves in Fig. 2. A step like signal of 80 mV amplitude @1 kHz is in- jected on the 10" input (top panel). The internal pairing of the FASP will produce in(9) + in(10) — out(10) and in(10) + in(11) — out(11). As in(9) = in(11) = 0 in our test, the amplitudes on out(10) and out(11) are equal and synchronous as can be seen in the middle pan- els of Fig. 2. The self and neighboring trigger mecha- nism is shown in the bottom panel. The signals labeled ”CS-Ch10” and ”CS-Ch11” are self-triggers on the output channels out(10) and out(11), with signal over threshold. Additionally, neighbor-triggers appear on the 9" and 12% channels, delayed wrt. to the self-triggers, which mark the digitization gate for under-threshold signals (if any). For completion the digital signals for the second order neigh- bors, the 8 and 13" channels, are also included to show that no signal is generated outside the requested range. Figure 2: Testing the basic features of FASP v03 on the test board during 1.2 us. From top to bottom: the input on pad entry 10, the summed analog output on channels 10 and 11 and the digital self-trigger signals on channels 8 — 13.
cessing of the chip are transparent from the curves in Fig. 2. A step like signal of 80 mV amplitude @1 kHz is in- jected on the 10" input (top panel). The internal pairing of the FASP will produce in(9) + in(10) — out(10) and in(10) + in(11) — out(11). As in(9) = in(11) = 0 in our test, the amplitudes on out(10) and out(11) are equal and synchronous as can be seen in the middle pan- els of Fig. 2. The self and neighboring trigger mecha- nism is shown in the bottom panel. The signals labeled ”CS-Ch10” and ”CS-Ch11” are self-triggers on the output channels out(10) and out(11), with signal over threshold. Additionally, neighbor-triggers appear on the 9" and 12% channels, delayed wrt. to the self-triggers, which mark the digitization gate for under-threshold signals (if any). For completion the digital signals for the second order neigh- bors, the 8 and 13" channels, are also included to show that no signal is generated outside the requested range. Figure 2: Testing the basic features of FASP v03 on the test board during 1.2 us. From top to bottom: the input on pad entry 10, the summed analog output on channels 10 and 11 and the digital self-trigger signals on channels 8 — 13.
Figure 3: The 6-FASPs FASPRO board (v02) prepared for mCBM integration; its back-side toward the detec- tor, housing the FASPs (top) and its front-side towards GETS board housing the Samtec™ ASP connector (mid- dle) and board-to-board connectors (left/right) (bottom pic- ture). Two FASP chips are mounted on the bottom row slots during local testings (top).
Figure 3: The 6-FASPs FASPRO board (v02) prepared for mCBM integration; its back-side toward the detec- tor, housing the FASPs (top) and its front-side towards GETS board housing the Samtec™ ASP connector (mid- dle) and board-to-board connectors (left/right) (bottom pic- ture). Two FASP chips are mounted on the bottom row slots during local testings (top).
For the FASP-FASP communication tests only two chips were mounted on the bottom row as presented in Fig. 3/top while the rest of FASP slots were left empty. In absence of the GETS board the FASPRO was operated by providing all LV values and clock from outside. The reading of the signals was done directly on the board using the oscillo- scope. Figure 4: Testing FASP-FASP communication on the FASPRO board from Fig. 3. The input on FAS P—0/ch— 15 (yellow) is seen on the output FAS P — 0/CH — 15 (green) and FAS P —1/CH —0 (blue). The FEE operates at 80 MHz clock rate (magenta).
For the FASP-FASP communication tests only two chips were mounted on the bottom row as presented in Fig. 3/top while the rest of FASP slots were left empty. In absence of the GETS board the FASPRO was operated by providing all LV values and clock from outside. The reading of the signals was done directly on the board using the oscillo- scope. Figure 4: Testing FASP-FASP communication on the FASPRO board from Fig. 3. The input on FAS P—0/ch— 15 (yellow) is seen on the output FAS P — 0/CH — 15 (green) and FAS P —1/CH —0 (blue). The FEE operates at 80 MHz clock rate (magenta).
scope. The input signal (yellow-top, label CJ) is injected on the 15" (last) channel of FASP — 0 (bottom left in Fig. 3). It generates the direct output (green signal - label C4 in figure) on the 15" output channel of FAS P — 0 but also the indirect, blue signal (labeled C2 in figure), regis- tered on the first output channel of FAS P — 1. As one can see, the chip-to-chip communication is proven as both signals are of equal amplitude (~ 1 V) without noise. A second order communication is also present in our FEE, the board-to-board one, which is designed over the left/right connectors on the front-side of FASPRO (see 3/bottom) not tested yet. Figure 5: A region on the back-side of the Bucharest- prototype to be installed at mCBM. A FASPRO board is also mounted on the FC connectors for the first mechanical integration tests.
scope. The input signal (yellow-top, label CJ) is injected on the 15" (last) channel of FASP — 0 (bottom left in Fig. 3). It generates the direct output (green signal - label C4 in figure) on the 15" output channel of FAS P — 0 but also the indirect, blue signal (labeled C2 in figure), regis- tered on the first output channel of FAS P — 1. As one can see, the chip-to-chip communication is proven as both signals are of equal amplitude (~ 1 V) without noise. A second order communication is also present in our FEE, the board-to-board one, which is designed over the left/right connectors on the front-side of FASPRO (see 3/bottom) not tested yet. Figure 5: A region on the back-side of the Bucharest- prototype to be installed at mCBM. A FASPRO board is also mounted on the FC connectors for the first mechanical integration tests.
Figure 1: Long run CR rate measurements using a TRD telescope operated with FASP/GETS FEE (top). Monitor- ing the DAQ synchronization and capturing the clock skew on the curve labeled ”TRD B row 0” on 6.09 at 11:00 AM local time (bottom). National Institute for Physics and Nuclear Engineering (IFIN-HH), Hadron Physics Department, Bucharest, Romania
Figure 1: Long run CR rate measurements using a TRD telescope operated with FASP/GETS FEE (top). Monitor- ing the DAQ synchronization and capturing the clock skew on the curve labeled ”TRD B row 0” on 6.09 at 11:00 AM local time (bottom). National Institute for Physics and Nuclear Engineering (IFIN-HH), Hadron Physics Department, Bucharest, Romania
Figure 2: The data flow of GETS for CR detection with emphasizes on different type of messages and their time correlation. an error code wrt. to ADC reading (here 0 no error). The message is followed by 9 other signal messages which are very similar to it as far as the t/ and data are concerned. This identifies a pulser event of synchronous, quasi-equal signal values in all channels (see Fig. 1 bottom). Since both FASP had data during current epoch, an epoch message is issued at the end (see label ’messEpoch”). Following the pulser signal the data flow contains a reading of 4 signal messages issued by FASP id = 0 on 4 adjacent channels and having the similar time-label t] = 124. This is identi- fied as a cluster produced by a CR. Since the cluster is fully contained in FASP 0 only its epoch message is send. The epoch counter of both FASPs is reset at the value 2097151 (to be compared with the epoch counter of the first CR hit 1860169 and last 130926). The ’super epoch message” is issued there and the ’super epoch counter” S Epoch is in- cremented. Last 6 messages are a second CR hit happening with 5 signal messages and the corresponding epoch mes- sage at the end.
Figure 2: The data flow of GETS for CR detection with emphasizes on different type of messages and their time correlation. an error code wrt. to ADC reading (here 0 no error). The message is followed by 9 other signal messages which are very similar to it as far as the t/ and data are concerned. This identifies a pulser event of synchronous, quasi-equal signal values in all channels (see Fig. 1 bottom). Since both FASP had data during current epoch, an epoch message is issued at the end (see label ’messEpoch”). Following the pulser signal the data flow contains a reading of 4 signal messages issued by FASP id = 0 on 4 adjacent channels and having the similar time-label t] = 124. This is identi- fied as a cluster produced by a CR. Since the cluster is fully contained in FASP 0 only its epoch message is send. The epoch counter of both FASPs is reset at the value 2097151 (to be compared with the epoch counter of the first CR hit 1860169 and last 130926). The ’super epoch message” is issued there and the ’super epoch counter” S Epoch is in- cremented. Last 6 messages are a second CR hit happening with 5 signal messages and the corresponding epoch mes- sage at the end.
The Self-triggered Pulse Amplification and Digitization asIC (SPADIC) [1] is the centrepiece of the Front-end Elec- tronic Boards (FEB) [2] and responsible for digitizing the analog signals coming from the detector. On a TRD mod- ule, up to 20 SPADICs are connected via e-links to a Read- out Board (ROB) which aggregates the data streams and sends them via optical links towards the Data Process- ing Board (DPB). The DPB processes the data of multiple ROBs by a high performance FPGA. Figure 1: Data acquisition chain of the final experiment along with the data format used at different stages.
The Self-triggered Pulse Amplification and Digitization asIC (SPADIC) [1] is the centrepiece of the Front-end Elec- tronic Boards (FEB) [2] and responsible for digitizing the analog signals coming from the detector. On a TRD mod- ule, up to 20 SPADICs are connected via e-links to a Read- out Board (ROB) which aggregates the data streams and sends them via optical links towards the Data Process- ing Board (DPB). The DPB processes the data of multiple ROBs by a high performance FPGA. Figure 1: Data acquisition chain of the final experiment along with the data format used at different stages.
Figure 2: Example of a hit message with 7 samples. Since version 2.2 the SPADIC encodes the data in 24-bit frames. message on one channel. Due to the geometry of the detec- tor chamber, the majority of physical hits will produce 3 to 4 spatially adjacent hit messages.
Figure 2: Example of a hit message with 7 samples. Since version 2.2 the SPADIC encodes the data in 24-bit frames. message on one channel. Due to the geometry of the detec- tor chamber, the majority of physical hits will produce 3 to 4 spatially adjacent hit messages.
For a configuration with 7 samples, this results in a total payload of 80 bit process can be su s per hit message. The feature extraction bdivided into separate stages (see Fig. 3). The first stage extracts the charge and refines the time in- formation for eac h hit message individually. After this step, assuming that, besides 4 bits for the channel number, only time and charge information with 12 bits each is forwarded, the payload is red uced to 28 bits corresponding to a reduc- tion of 65 %. At the second stage a cluster finding algo- rithm selects all hit messages which most likely belong to a single physical hit. Figure 3: The two separate stages of the feature extraction process. In this example a hit resulting in three hit mes- sages is reconstructed.
For a configuration with 7 samples, this results in a total payload of 80 bit process can be su s per hit message. The feature extraction bdivided into separate stages (see Fig. 3). The first stage extracts the charge and refines the time in- formation for eac h hit message individually. After this step, assuming that, besides 4 bits for the channel number, only time and charge information with 12 bits each is forwarded, the payload is red uced to 28 bits corresponding to a reduc- tion of 65 %. At the second stage a cluster finding algo- rithm selects all hit messages which most likely belong to a single physical hit. Figure 3: The two separate stages of the feature extraction process. In this example a hit resulting in three hit mes- sages is reconstructed.
As a basis for the simulation, an intrinsic energy resolu- tion of the detector of 10 % and a noise level of gapc = 1 was assumed. Figure 4 shows different methods used to extract the signal height of the simulated signals. Figure 4: Comparison of the performance of different fea- ture extraction methods.
As a basis for the simulation, an intrinsic energy resolu- tion of the detector of 10 % and a noise level of gapc = 1 was assumed. Figure 4 shows different methods used to extract the signal height of the simulated signals. Figure 4: Comparison of the performance of different fea- ture extraction methods.
Figure 5: Influence of noise on the max-ADC method.
Figure 5: Influence of noise on the max-ADC method.
Figure 2: Sketch of the setup, showing relative distances. The inner red box marks the position of the GIF !°’Cs ir- radiator, while the ~z beam is sketched as a black line. Figure 1: Tvo TRD MWPCs (arrow) installed in the down- stream field of the GIF irradiator, and in-line with the Lt beam from CERN-SPS/T2/H4.
Figure 2: Sketch of the setup, showing relative distances. The inner red box marks the position of the GIF !°’Cs ir- radiator, while the ~z beam is sketched as a black line. Figure 1: Tvo TRD MWPCs (arrow) installed in the down- stream field of the GIF irradiator, and in-line with the Lt beam from CERN-SPS/T2/H4.
In October 2018, measurements with CBM-TRD MWPCs have been conducted at the CERN Gamma Irra- diation Facility (GIF**) [1], during which the detectors have been exposed to the y flux of a 14TBq 1°"Cs source and 1 beam at the same time. Two modules were aligned in 1:1 coverage with respect to the passing js beam. As yp ref- erence counter, two scintillators of the GIF have been used, which are permanently installed outside the GIF cave and thus are shielded from the GIF source irradiation field, but passed by the ys beam only. A third scintillator has been in- stalled directly in front of the TRD MWPCs, matching the active (recorded) area of the tested detectors. A sketch of the different detectors is shown in Figure 2.
In October 2018, measurements with CBM-TRD MWPCs have been conducted at the CERN Gamma Irra- diation Facility (GIF**) [1], during which the detectors have been exposed to the y flux of a 14TBq 1°"Cs source and 1 beam at the same time. Two modules were aligned in 1:1 coverage with respect to the passing js beam. As yp ref- erence counter, two scintillators of the GIF have been used, which are permanently installed outside the GIF cave and thus are shielded from the GIF source irradiation field, but passed by the ys beam only. A third scintillator has been in- stalled directly in front of the TRD MWPCs, matching the active (recorded) area of the tested detectors. A sketch of the different detectors is shown in Figure 2.
Figure 3: Measured HV anode and drift currents wrt. counted hit-rates per per pad (1cm?). Anode voltage of 1850 V and drift voltage of 500 V. The current scaling is found to follow a linear correlation up to the highest mea- sured detector load. Figure 3 shows anode and drift currents as measured with the supplying HV units in dependence of the mean counting rates per pad, which corresponds to a counting rate per 1 cm? of detector surface. The found current scal- ing is compatible with a linear description, which is inter- preted as stable detector gain in the tested detector loads.
Figure 3: Measured HV anode and drift currents wrt. counted hit-rates per per pad (1cm?). Anode voltage of 1850 V and drift voltage of 500 V. The current scaling is found to follow a linear correlation up to the highest mea- sured detector load. Figure 3 shows anode and drift currents as measured with the supplying HV units in dependence of the mean counting rates per pad, which corresponds to a counting rate per 1 cm? of detector surface. The found current scal- ing is compatible with a linear description, which is inter- preted as stable detector gain in the tested detector loads.
energy deposits based on °°Fe decay BR and X - ArCO» interaction. We have considered the following Fe lines Ka = 5.895 keV (89%), Kg = 6.492 keV (11%). Consid- ering only the photoelectric effect on the Ar K-shell (bind- ing energy 3.2 keV) and the Auger electron of 2.7 keV and emission probability of 88% one can model a 4 spec- tral components with the following energies and theoret- ical abundances: 5.992 keV (9.7%), 5.395 keV (9.7%), 3.292 keV (1.2%) and 2.695 keV (9.8%). The spectral de- composition of the reconstructed energy spectrum is pre- sented in the figure under the label ’Model components”; the 5.992 keV component in red and the 5.395 keV com- ponent in blue. The sum of the 4 components is shown in green under the label ’Model 4Gauss”. The difference between the reconstructed spectrum and the theoretical 4 components spectrum is shown as ’Residual spectrum’. Figure 1: Spectral deconvolution of the °°Fe spectrum us- ing four components for selected clusters based on readout topology (top). Energy calibration for a range of 5 anode voltages (bottom).
energy deposits based on °°Fe decay BR and X - ArCO» interaction. We have considered the following Fe lines Ka = 5.895 keV (89%), Kg = 6.492 keV (11%). Consid- ering only the photoelectric effect on the Ar K-shell (bind- ing energy 3.2 keV) and the Auger electron of 2.7 keV and emission probability of 88% one can model a 4 spec- tral components with the following energies and theoret- ical abundances: 5.992 keV (9.7%), 5.395 keV (9.7%), 3.292 keV (1.2%) and 2.695 keV (9.8%). The spectral de- composition of the reconstructed energy spectrum is pre- sented in the figure under the label ’Model components”; the 5.992 keV component in red and the 5.395 keV com- ponent in blue. The sum of the 4 components is shown in green under the label ’Model 4Gauss”. The difference between the reconstructed spectrum and the theoretical 4 components spectrum is shown as ’Residual spectrum’. Figure 1: Spectral deconvolution of the °°Fe spectrum us- ing four components for selected clusters based on readout topology (top). Energy calibration for a range of 5 anode voltages (bottom).
Due to various systematic effects in cluster reconstruc- tion, estimations as presented in Fig. 1 (bottom) were re- peated for all 9 anode wires serving one pad row. The val- ues for the slope obtained for each pair (Uanode, [Danode) (excepting the anode wires which are on the row bound- aries), are represented in Fig. 2 (after a common offset was forced on all energy calibration fits) as function of the Uanode- For each I Danode an exponential fit is performed. Excepting two outliers, all exponential model describing the gain dependence on the applied polarization have simi- lar parameters thus proving the method consistency. Figure 2: The exponential model (continuous lines) de- scribing the TRD gas gain dependency on the anode volt- age fit through the energy calibration slopes of 7 out of 9 anode wires.
Due to various systematic effects in cluster reconstruc- tion, estimations as presented in Fig. 1 (bottom) were re- peated for all 9 anode wires serving one pad row. The val- ues for the slope obtained for each pair (Uanode, [Danode) (excepting the anode wires which are on the row bound- aries), are represented in Fig. 2 (after a common offset was forced on all energy calibration fits) as function of the Uanode- For each I Danode an exponential fit is performed. Excepting two outliers, all exponential model describing the gain dependence on the applied polarization have simi- lar parameters thus proving the method consistency. Figure 2: The exponential model (continuous lines) de- scribing the TRD gas gain dependency on the anode volt- age fit through the energy calibration slopes of 7 out of 9 anode wires.
Figure 3: Reconstructed °°Fe spectrum measured at 2 kV anode voltage for the four most central anodes and various measured statistics of yields for the spectral components as compared with the theoretical values (green).
Figure 3: Reconstructed °°Fe spectrum measured at 2 kV anode voltage for the four most central anodes and various measured statistics of yields for the spectral components as compared with the theoretical values (green).
Figure 1: A picture of the experimental setup. The setup used for the investigations (see Fig. 1) con- sisted of an acrylic hood, enabling the creation of a humidity-controlled atmosphere around the TRD module. The module and parts of the supplying detector gas lines were positioned inside this volume. The ambient humid- ity inside the hood was influenced using a dry air supply. Every part of the tubing outside the hood and upstream of the sensors was made of steel. The ambient temperature and humidity were logged, as was the flow of the used gas (Sagox 18: Ar/CO2 (82/18)), the humidity and oxygen con- centration inside the gas system and the pressure difference between in- and outside of the system. The fraction of humidity induced by the polyamide hoses is measured right before and right after a measurement with
Figure 1: A picture of the experimental setup. The setup used for the investigations (see Fig. 1) con- sisted of an acrylic hood, enabling the creation of a humidity-controlled atmosphere around the TRD module. The module and parts of the supplying detector gas lines were positioned inside this volume. The ambient humid- ity inside the hood was influenced using a dry air supply. Every part of the tubing outside the hood and upstream of the sensors was made of steel. The ambient temperature and humidity were logged, as was the flow of the used gas (Sagox 18: Ar/CO2 (82/18)), the humidity and oxygen con- centration inside the gas system and the pressure difference between in- and outside of the system. The fraction of humidity induced by the polyamide hoses is measured right before and right after a measurement with
Subsequent to this first decrease, the ambient humidity was reduced rapidly. After this reduction, the humidity in the gas system is around 1150 ppmV (after 400 h of mea- surement time, see Fig. 2). This decrease of humidity is due to a drying of the chamber, since all materials without the chamber have been shown to be constant under change of ambient humidity before. Figure 2: The development of the humidity concentration inside the gas system after drying the acrylic hood, begin- ning at 186h measurement time, and the ambient relative humidity.
Subsequent to this first decrease, the ambient humidity was reduced rapidly. After this reduction, the humidity in the gas system is around 1150 ppmV (after 400 h of mea- surement time, see Fig. 2). This decrease of humidity is due to a drying of the chamber, since all materials without the chamber have been shown to be constant under change of ambient humidity before. Figure 2: The development of the humidity concentration inside the gas system after drying the acrylic hood, begin- ning at 186h measurement time, and the ambient relative humidity.
Figure 3: The development of humidity and oxygen con- centration and the ambient gas flow.
Figure 3: The development of humidity and oxygen con- centration and the ambient gas flow.
The TRD will deliver particle identification crucial for the investigation of rare dilepton channels and will be able to separate charge states for the hypernuclei program. For the best understanding of the detector it is necessary to sim- ulate all effects as realistic as possible, in order to get the best overview of all expected performances and possible is- sues. By implementing the front-end electronic algorithms in the simulation it is now possible to study all detector as- pects in more detail. Since the Monte-Carlo information generated by Geant only contains the total charge deposi- tion, but does not calculate the ionization at specific po- sitions inside the gas volume, this has to included in the digitization step, e.g. to account for statistical fluctuations in the ionization process. The charge is now distributed ac- cording to the Bethe-Bloch curve for Xe [2] (see Fig. 1) to calculate the integrated probability for an ionization to happen at a certain distance using Eqs. | and 2. Figure 1: Relative energy loss in Xe [2]. Afterwards, the charge is distributed to different read out channels via the Pad-Response-Function (PRF) based on the Mathieson formula (see Fig. 2). With these steps we get multiple little ionizations at different positions along the track of the particle, together with their induced charges on each read out channel around its original position. In reality the ion drifts, which are collected by the front-end electronics, are formed into pulse shapes corresponding to the shaper function in Eq. 3 [1],
The TRD will deliver particle identification crucial for the investigation of rare dilepton channels and will be able to separate charge states for the hypernuclei program. For the best understanding of the detector it is necessary to sim- ulate all effects as realistic as possible, in order to get the best overview of all expected performances and possible is- sues. By implementing the front-end electronic algorithms in the simulation it is now possible to study all detector as- pects in more detail. Since the Monte-Carlo information generated by Geant only contains the total charge deposi- tion, but does not calculate the ionization at specific po- sitions inside the gas volume, this has to included in the digitization step, e.g. to account for statistical fluctuations in the ionization process. The charge is now distributed ac- cording to the Bethe-Bloch curve for Xe [2] (see Fig. 1) to calculate the integrated probability for an ionization to happen at a certain distance using Eqs. | and 2. Figure 1: Relative energy loss in Xe [2]. Afterwards, the charge is distributed to different read out channels via the Pad-Response-Function (PRF) based on the Mathieson formula (see Fig. 2). With these steps we get multiple little ionizations at different positions along the track of the particle, together with their induced charges on each read out channel around its original position. In reality the ion drifts, which are collected by the front-end electronics, are formed into pulse shapes corresponding to the shaper function in Eq. 3 [1],
The calculation of the actual pulses and the conversion of their charge values into ADC channels allows for real- istic investigations of energy resolutions with different re- construction methods, precise studies of trigger behaviour and better handling of correlated and uncorrelated noise, as well as crosstalk between the channels. The resulting simulated pulses are shown in Fig. 3. Figure 3: Self-triggered pulses for different ranges of en- ergy deposition in the detector, simulated using the SIS100 electron setup and UrQMD events at 8 A GeV.
The calculation of the actual pulses and the conversion of their charge values into ADC channels allows for real- istic investigations of energy resolutions with different re- construction methods, precise studies of trigger behaviour and better handling of correlated and uncorrelated noise, as well as crosstalk between the channels. The resulting simulated pulses are shown in Fig. 3. Figure 3: Self-triggered pulses for different ranges of en- ergy deposition in the detector, simulated using the SIS100 electron setup and UrQMD events at 8 A GeV.
Figure 2: Measured charge distribution of a charged par- ticle passing through the detector and creating a signal on three adjacent readout channels [1]. Institut fiir Kernphysik, Frankfurt am Main, Germany
Figure 2: Measured charge distribution of a charged par- ticle passing through the detector and creating a signal on three adjacent readout channels [1]. Institut fiir Kernphysik, Frankfurt am Main, Germany
Figure 1: Example of modeling the response of BuchTRD+FASP for the mCBM set-up; physical signal fronts (black markers), amplified/shaped signals generation (thin black) and FASP analog output (thin red) on two ad- jacent channels (12 and 13) out of 2880.
Figure 1: Example of modeling the response of BuchTRD+FASP for the mCBM set-up; physical signal fronts (black markers), amplified/shaped signals generation (thin black) and FASP analog output (thin red) on two ad- jacent channels (12 and 13) out of 2880.
black. The FASP amplified/shaped response is calculated based on convoluted singled value CADENCE estimates and marked by the thin continuous black curve. The ana- log output of FASP is marked in thin red, where the flat top marks the ADC converted value of the input charge and the small indentation, before the plateau, the moment of self- trigger. A complex read-out structure, generated by pile-up (more than three signal fronts), is shown in the same fig- ure at e.g. time & 8000 clk (marked by red arrow). The shaped FASP signal, especially for the channel 12 (top), becomes wriggled, the self-trigger is delayed wrt. prompt signal and its amplitude lowered, both design features of the ASIC in pile-up conditions, correctly reproduced by the simulation. Figure 2: FASP-digits representation of TRD response to top rate interactions in the mCBM set-up. The time on the horizontal axis is measured in FASP clocks of 12.5 ns.
black. The FASP amplified/shaped response is calculated based on convoluted singled value CADENCE estimates and marked by the thin continuous black curve. The ana- log output of FASP is marked in thin red, where the flat top marks the ADC converted value of the input charge and the small indentation, before the plateau, the moment of self- trigger. A complex read-out structure, generated by pile-up (more than three signal fronts), is shown in the same fig- ure at e.g. time & 8000 clk (marked by red arrow). The shaped FASP signal, especially for the channel 12 (top), becomes wriggled, the self-trigger is delayed wrt. prompt signal and its amplitude lowered, both design features of the ASIC in pile-up conditions, correctly reproduced by the simulation. Figure 2: FASP-digits representation of TRD response to top rate interactions in the mCBM set-up. The time on the horizontal axis is measured in FASP clocks of 12.5 ns.
Even though a pre mass-production for the TOF FAIR phase 0 program of about 150 MRPC3a/b counter is fin- ished [3, 4, 5] the MRPC R&D especially for the inner TOF region is still ongoing. The development of the ce- ramic single cell MRPCs, foreseen for the Beam Fragmen- tation TO Counter (BFTC), is well progressed and test beam results are presented here [6]. The R&D for this type of counter will continue till 2021 and will be finalized based on test beam results at Rossendorf and mCBM. In the same time frame the PDR of the inner TOF wall is intended. The R&D regarding electrode materials and final design of this counter type (MRPC La-c) is ongoing aiming at tests of the latest prototype [7] in the mCBM beam time 2019. The counter R&D for the intermediate and outer region is mostly concentrated on the HV electrode and the FDR will be based on the mCBM beam time in 2020. Figure 1: mTOF wall at mCBM (left plot) and eTOF wheel at STAR (right plot).
Even though a pre mass-production for the TOF FAIR phase 0 program of about 150 MRPC3a/b counter is fin- ished [3, 4, 5] the MRPC R&D especially for the inner TOF region is still ongoing. The development of the ce- ramic single cell MRPCs, foreseen for the Beam Fragmen- tation TO Counter (BFTC), is well progressed and test beam results are presented here [6]. The R&D for this type of counter will continue till 2021 and will be finalized based on test beam results at Rossendorf and mCBM. In the same time frame the PDR of the inner TOF wall is intended. The R&D regarding electrode materials and final design of this counter type (MRPC La-c) is ongoing aiming at tests of the latest prototype [7] in the mCBM beam time 2019. The counter R&D for the intermediate and outer region is mostly concentrated on the HV electrode and the FDR will be based on the mCBM beam time in 2020. Figure 1: mTOF wall at mCBM (left plot) and eTOF wheel at STAR (right plot).
wheel was build comprising 36 mod ules arranged in 12 sec- tors contributing to the STAR detector upgrade program. The 108 counters were read out by our current CBM stan- dard free streaming DAQ based on the AFCK and portions of data selected based on the STAR trigger token were send to the STAR event builder. That this approach was success- ful is demonstrated in Fig. 2 [9]. shows the PID capability of eTOF lower row show the phase space d n The upper row of plots while the plots of the istribution of the vari- ous hadrons. The data providing this plots were obtained during the beginning of the Beam Energy Scan campaign (BESID in 2019. Dedicated data analysis show a system time resolution in the order of 85 separation of kaons from pions up ps [10] which allows a to 1.8 GeV/c. In order to reach such PID quality a sophisticated calibration is re- quired [11]. Figure 2: Particle identification capability of eTOF demon- strated at STAR (upper plots) and phase space distributions for pions, kaons and protons (lower plots).
wheel was build comprising 36 mod ules arranged in 12 sec- tors contributing to the STAR detector upgrade program. The 108 counters were read out by our current CBM stan- dard free streaming DAQ based on the AFCK and portions of data selected based on the STAR trigger token were send to the STAR event builder. That this approach was success- ful is demonstrated in Fig. 2 [9]. shows the PID capability of eTOF lower row show the phase space d n The upper row of plots while the plots of the istribution of the vari- ous hadrons. The data providing this plots were obtained during the beginning of the Beam Energy Scan campaign (BESID in 2019. Dedicated data analysis show a system time resolution in the order of 85 separation of kaons from pions up ps [10] which allows a to 1.8 GeV/c. In order to reach such PID quality a sophisticated calibration is re- quired [11]. Figure 2: Particle identification capability of eTOF demon- strated at STAR (upper plots) and phase space distributions for pions, kaons and protons (lower plots).
Since the signal from the chambers of the mini-module is differentiated and shaped to have a pulse length not more than 5 ns for operation under high counting rates, the re- sults of 2017 beam-tests lead us to the assumption, that the tested and tl issue TDCs are not sensitive to very short signa his causes the loss of efficiency. In order s (< 5 ns) to fix this we implemented an external board for signal stretch- ing. The outputs of this stretcher increases the d uration of the input signal by 3.7 ns. To test this solution, a beam session was conducted at the ELBE accelerator (HZDR) in April 2018. Figure 1: Efficiency as function of electric field strength for three RPC cells and different types of used readout elec- tronics.
Since the signal from the chambers of the mini-module is differentiated and shaped to have a pulse length not more than 5 ns for operation under high counting rates, the re- sults of 2017 beam-tests lead us to the assumption, that the tested and tl issue TDCs are not sensitive to very short signa his causes the loss of efficiency. In order s (< 5 ns) to fix this we implemented an external board for signal stretch- ing. The outputs of this stretcher increases the d uration of the input signal by 3.7 ns. To test this solution, a beam session was conducted at the ELBE accelerator (HZDR) in April 2018. Figure 1: Efficiency as function of electric field strength for three RPC cells and different types of used readout elec- tronics.
sie, Pierre value are typical for this type of detector. It does not have spikes or gaps, which ensures the impedance matching of the connection mini-module to PADI. Figure 2: ToT spectra (left row) and ToT-to-time correla- tions (right row) of channel 4 for different values of the discrimination threshold indicated on the left. Minimodule is connected to the TDC via a signal stretcher. All time scales in ns.
sie, Pierre value are typical for this type of detector. It does not have spikes or gaps, which ensures the impedance matching of the connection mini-module to PADI. Figure 2: ToT spectra (left row) and ToT-to-time correla- tions (right row) of channel 4 for different values of the discrimination threshold indicated on the left. Minimodule is connected to the TDC via a signal stretcher. All time scales in ns.
connection format to the NINO board format was made. Such an adapter is not resistant to noise interference and should not be used for accurate timing measurements. The results of the efficiency measurements and comparison to those obtained with PADI and MAXIM electronics are shown in Fig. 4. The results for NINO agree well with the MAXIM 3760 results. It should be noticed, that dur- ing this measurements due to high noise level the threshold value was set almost twice as high as that used in ALICE measurements. Furthermore, the width of the ToT spec- trum, which is shown in Fig. 5, exactly matched to the pulse length of 5 ns. In addition, the shape of ToT spectra agrees well with spectra of the ALICE TOF strip RPCs, obtained using the same electronics during ALICE R&D [4]. The time resolution, shown on Fig. 6 as function of the field strength in the gaps, amounts to 87 ps across the working plateau of the chamber. Figure 4: Registration efficiency versus field strength in a chamber for connecting discrete electronics based on a MAXIM 3760 amplifier (magenta curve), NINO elec- tronics (blue and red curves) and PADI electronics (black curve)
connection format to the NINO board format was made. Such an adapter is not resistant to noise interference and should not be used for accurate timing measurements. The results of the efficiency measurements and comparison to those obtained with PADI and MAXIM electronics are shown in Fig. 4. The results for NINO agree well with the MAXIM 3760 results. It should be noticed, that dur- ing this measurements due to high noise level the threshold value was set almost twice as high as that used in ALICE measurements. Furthermore, the width of the ToT spec- trum, which is shown in Fig. 5, exactly matched to the pulse length of 5 ns. In addition, the shape of ToT spectra agrees well with spectra of the ALICE TOF strip RPCs, obtained using the same electronics during ALICE R&D [4]. The time resolution, shown on Fig. 6 as function of the field strength in the gaps, amounts to 87 ps across the working plateau of the chamber. Figure 4: Registration efficiency versus field strength in a chamber for connecting discrete electronics based on a MAXIM 3760 amplifier (magenta curve), NINO elec- tronics (blue and red curves) and PADI electronics (black curve)
Figure 3: Typical MIP signal from the RPC of the mini- module.
Figure 3: Typical MIP signal from the RPC of the mini- module.
Figure 5: Tot-to-Time correlation (left row) and ToT spec- tra (middle row) of the BFTC mini-module in compari- son to the pulse-width spectra of ALICE strip RPCs ob- tained with NINO electronics for different values of the field strength in the gas gap (applied voltage in the case of ALICE data). Figure 6: Time resolution of the mini-module RPC cell number 4 in case of using NINO electronics as function of the field strength in the gas gap.
Figure 5: Tot-to-Time correlation (left row) and ToT spec- tra (middle row) of the BFTC mini-module in compari- son to the pulse-width spectra of ALICE strip RPCs ob- tained with NINO electronics for different values of the field strength in the gas gap (applied voltage in the case of ALICE data). Figure 6: Time resolution of the mini-module RPC cell number 4 in case of using NINO electronics as function of the field strength in the gas gap.
Figure 1: The two identical MGMSRPCs with 100 Ohm impedance mounted on the back pannel of the housing box. A MGMSRPC prototype with 32 readout channels was designed and built as basic cell for the inner zone of the CBM-TOF wall. The aim is to exploit efficiently both the active area of the detector as well as the 32 channels of a single front-end electronics baseboard for CBM-TOF. The inner architecture of the detector is the same as for RPC2015DS prototype [1], successfully tested in two in- beam test campaignes in both triggered [2] and free stream- ing mode operation [3]. The strip width of the readout elec- trode was obtained using the method described in [4]. A readout strip width of 1.27 mm corresponds to 100 © char- acteristic impedance of the differential transmission line for the present architecture. The obtained readout strip width is very close to the value of the strip width of RPC2015DS. However, in order to exploit efficiently the maximum avail- able size of the low resistivity glass, the strip width of the high voltage electrode increased from 5.6 mm to 7.37 mm. Consequently, the strip pitch, the same for both HV and readout electrodes, increased from 7.2 mm to 9.02 mm. The 60 mm strip length corresponds to the counters of the CBM-TOF inner zone with the highest granularity. After assembling, the two prototypes were preliminary tested in the detector laboratory. A pulser signal was split in two, one of the signals being used to trigger the oscilloscope, the other one being split again, one of these being inverted. The negative signal was injected in the anode and the positive one in the cathode readout strips, at one end, and recorded at the other end directly on the oscilloscope. The measured propagated signals through the transmission line did not re- veal any significant reflections. A cross talk test was also performed, injecting both polarity signals on the main strip (CH1 and CH2) and recording the signals of the first neigh- bour strip (CH3 and CH4), at the same side where the sig- nal was injected, the oposite side being closed and matched to 50 Ohm. As can be observed in the screen capture shown in Fig.2, the cross talk is negligible.
Figure 1: The two identical MGMSRPCs with 100 Ohm impedance mounted on the back pannel of the housing box. A MGMSRPC prototype with 32 readout channels was designed and built as basic cell for the inner zone of the CBM-TOF wall. The aim is to exploit efficiently both the active area of the detector as well as the 32 channels of a single front-end electronics baseboard for CBM-TOF. The inner architecture of the detector is the same as for RPC2015DS prototype [1], successfully tested in two in- beam test campaignes in both triggered [2] and free stream- ing mode operation [3]. The strip width of the readout elec- trode was obtained using the method described in [4]. A readout strip width of 1.27 mm corresponds to 100 © char- acteristic impedance of the differential transmission line for the present architecture. The obtained readout strip width is very close to the value of the strip width of RPC2015DS. However, in order to exploit efficiently the maximum avail- able size of the low resistivity glass, the strip width of the high voltage electrode increased from 5.6 mm to 7.37 mm. Consequently, the strip pitch, the same for both HV and readout electrodes, increased from 7.2 mm to 9.02 mm. The 60 mm strip length corresponds to the counters of the CBM-TOF inner zone with the highest granularity. After assembling, the two prototypes were preliminary tested in the detector laboratory. A pulser signal was split in two, one of the signals being used to trigger the oscilloscope, the other one being split again, one of these being inverted. The negative signal was injected in the anode and the positive one in the cathode readout strips, at one end, and recorded at the other end directly on the oscilloscope. The measured propagated signals through the transmission line did not re- veal any significant reflections. A cross talk test was also performed, injecting both polarity signals on the main strip (CH1 and CH2) and recording the signals of the first neigh- bour strip (CH3 and CH4), at the same side where the sig- nal was injected, the oposite side being closed and matched to 50 Ohm. As can be observed in the screen capture shown in Fig.2, the cross talk is negligible.
Figure 2: CH1 and CH2 - injected signals; CH3 and CH4 neglijible cross-talk in the first neighbour.
Figure 2: CH1 and CH2 - injected signals; CH3 and CH4 neglijible cross-talk in the first neighbour.
Figure 3: Snap-shot of real cosmic ray signals. We conditioned the detector applying gradually the high voltage up to the nominal value (2 x 5.5 kV). Real signals produced by cosmic rays were recorded. A snapshot of the signals from both cathode (CH1) and anode (CH2) readout strips recorded directly from the detector at one end of the strip and the output of the NINO FEE [5] i.e the LVDS signal (CH4) and the logic OR (CH3), recorded at the other end, are presented in Fig.3. As can be seen, one does not observe reflections on a time scale of 200 ns (20 ns/div).
Figure 3: Snap-shot of real cosmic ray signals. We conditioned the detector applying gradually the high voltage up to the nominal value (2 x 5.5 kV). Real signals produced by cosmic rays were recorded. A snapshot of the signals from both cathode (CH1) and anode (CH2) readout strips recorded directly from the detector at one end of the strip and the output of the NINO FEE [5] i.e the LVDS signal (CH4) and the logic OR (CH3), recorded at the other end, are presented in Fig.3. As can be seen, one does not observe reflections on a time scale of 200 ns (20 ns/div).
x? limit. With an 8 % maximum chance of a mismatch in a multi-track environment (left plot) compared to a clean single-track reference scenario (right plot; obtained with a muon generator) this method bias needs to be taken into account for efficiency determination in the data analysis. Figure 2: Matching acceptance (blue curve) and purity (red curve) of DUT hits with selector tracks as a function of the (reduced) \? distance measure limit in a multi-track (left plot) and in a single-track (right plot) environment. As a point of reference, the y? acceptance of a purely Gaussian response function with 3 degrees of freedom (black curve) is given.
x? limit. With an 8 % maximum chance of a mismatch in a multi-track environment (left plot) compared to a clean single-track reference scenario (right plot; obtained with a muon generator) this method bias needs to be taken into account for efficiency determination in the data analysis. Figure 2: Matching acceptance (blue curve) and purity (red curve) of DUT hits with selector tracks as a function of the (reduced) \? distance measure limit in a multi-track (left plot) and in a single-track (right plot) environment. As a point of reference, the y? acceptance of a purely Gaussian response function with 3 degrees of freedom (black curve) is given.
Figure 1: Particle species vs. production process of all (left plot) simulated charged particles created in the target vol- ume which traverse all tracking stations and of those only which a selector track has been matched to (right plot).
Figure 1: Particle species vs. production process of all (left plot) simulated charged particles created in the target vol- ume which traverse all tracking stations and of those only which a selector track has been matched to (right plot).
mTOF is one of the sub-systems in the mCBM experi- ment [1], designed to measure the time of flight of charged particles. It features 5 modules (cf. Fig. 1 right) cov- ering an total area of 1.25 m x 1.5 m. During the com- missioning beam-time in December 2018 the mTOF wall was rearranged in order to allow quality assurance of the detectors. Therefore 2 stacks of 2 and 3 modules each (as shown in Fig. 1) were positioned at an angle of 25° with respect to the beam axis and 234 cm from the target. Each module contains five high rate MRPCs called MRPC3a [2], which contains low resistive glass sheets [3]. The high rate capability and good time resolution of these detectors are crucial for particle identification in high rate experiments. The first stack (module 4 and 5 in Fig. 1) of mTOF con- tains two modules, overlapping in the z-axis direction. It is located closer to the beam pipe and therefore exposed to higher particle fluxes. With this stack the aim is to test the rate capability of the MRPC3a counters. However, it is not trivial to get the efficiency of the counters reliably well with only one additional reference. Therefore the information of a TO (diamond counter) is needed. The situation is better for the second stack with three modules (module 1,2 and 3) overlapping in the z-axis direction. With tracks travers- ing all three modules, the efficiency of the MRPCs can be obtained more easily due to one more layer as reference detector. Figure 1: The mTOF experimental set-up and module structure. Left: The mTOF experimental set-up. It is lo- cated 234 cm away from the target, positioned under an angle of 25° with respect to primary beam. Right: One mTOF module. Each module contains five MRPC3as.
mTOF is one of the sub-systems in the mCBM experi- ment [1], designed to measure the time of flight of charged particles. It features 5 modules (cf. Fig. 1 right) cov- ering an total area of 1.25 m x 1.5 m. During the com- missioning beam-time in December 2018 the mTOF wall was rearranged in order to allow quality assurance of the detectors. Therefore 2 stacks of 2 and 3 modules each (as shown in Fig. 1) were positioned at an angle of 25° with respect to the beam axis and 234 cm from the target. Each module contains five high rate MRPCs called MRPC3a [2], which contains low resistive glass sheets [3]. The high rate capability and good time resolution of these detectors are crucial for particle identification in high rate experiments. The first stack (module 4 and 5 in Fig. 1) of mTOF con- tains two modules, overlapping in the z-axis direction. It is located closer to the beam pipe and therefore exposed to higher particle fluxes. With this stack the aim is to test the rate capability of the MRPC3a counters. However, it is not trivial to get the efficiency of the counters reliably well with only one additional reference. Therefore the information of a TO (diamond counter) is needed. The situation is better for the second stack with three modules (module 1,2 and 3) overlapping in the z-axis direction. With tracks travers- ing all three modules, the efficiency of the MRPCs can be obtained more easily due to one more layer as reference detector. Figure 1: The mTOF experimental set-up and module structure. Left: The mTOF experimental set-up. It is lo- cated 234 cm away from the target, positioned under an angle of 25° with respect to primary beam. Right: One mTOF module. Each module contains five MRPC3as.
mTOF beam test The free-streaming mTOF readout chain is composed of the PADI1O preamp Digital Converter (1 lifier, the self-triggered GET4 Time to [DC), the GBTX based readout board (ROB) and the FPGA based Data Processing Board (DPB) AFCK as shown in to the FLES (a hig located in the Green Fig. 2. The data from mTOF are sent h-performance online computer farm) IT Cube where they are combined with the data from other sub systems and stored to a file system. The so called tsa-fi es which contain the raw information were unpacked during the December beam-time for TOF the first time in the FAIR MQ framework producing a root file which can be analyzed by the CBM root framework. The performance of mTOF was firstly tested in beam during mCBM commissioning run in Dec. 2018 carried out at SIS18 in GSI. The beam was generated by 1.2 AGeV Ag hitting 10% Au target, whose total thickness is 0.1 cm with 0.5 cm in diameter. The working gas mixture for mTOF is 90 % CoHoeFy (R134a), 5 % i-C4H 9 and 5 % SFg. Due to a malfunction of the gas distribution panel, the five mod- ules had to be daisy-chained, flowed by the gas mixture for 15 days at the speed of only 93 ml/min for the beam test. It turned out that this was not sufficient to have the same gas quality in all modules. Consequently, the high voltage (HV) of five modules could be set to 4 t 5300 V, 4 t 5300 V,
mTOF beam test The free-streaming mTOF readout chain is composed of the PADI1O preamp Digital Converter (1 lifier, the self-triggered GET4 Time to [DC), the GBTX based readout board (ROB) and the FPGA based Data Processing Board (DPB) AFCK as shown in to the FLES (a hig located in the Green Fig. 2. The data from mTOF are sent h-performance online computer farm) IT Cube where they are combined with the data from other sub systems and stored to a file system. The so called tsa-fi es which contain the raw information were unpacked during the December beam-time for TOF the first time in the FAIR MQ framework producing a root file which can be analyzed by the CBM root framework. The performance of mTOF was firstly tested in beam during mCBM commissioning run in Dec. 2018 carried out at SIS18 in GSI. The beam was generated by 1.2 AGeV Ag hitting 10% Au target, whose total thickness is 0.1 cm with 0.5 cm in diameter. The working gas mixture for mTOF is 90 % CoHoeFy (R134a), 5 % i-C4H 9 and 5 % SFg. Due to a malfunction of the gas distribution panel, the five mod- ules had to be daisy-chained, flowed by the gas mixture for 15 days at the speed of only 93 ml/min for the beam test. It turned out that this was not sufficient to have the same gas quality in all modules. Consequently, the high voltage (HV) of five modules could be set to 4 t 5300 V, 4 t 5300 V,
Good quality data for mTOF was obtained during almost the whole beam test. The first and preliminary analysis re- sults of mTOF were obtained for RUN 49 were all partic- ipating sub-systems were combined. Fig. 3 shows the av- erage cluster size (ACS) for all 25 counters (grouped in 5 modules with different colors). The average of each mod- ule is indicated by the black dashed line. It decreases from about 1.2 to 1.1 depending on the applied HV which is in all cases kind of small due to the insufficient applied work- ing voltage. In addition a trend of the average cluster size is seen in each module. The MRPCs located on the top of the “hanging” module shows the largest ACS. The further down the counter is, the lower the ACS. The most prob- able reason for this observation could be the temperature gradient but also the pressure gradient in the module. Figure 3: The average cluster size of each counter in mTOF. Each point represents a counter. The gas flowed through counters from counter | to counter 25 in sequence. The dotted line is the average cluster size of five counters in each module.
Good quality data for mTOF was obtained during almost the whole beam test. The first and preliminary analysis re- sults of mTOF were obtained for RUN 49 were all partic- ipating sub-systems were combined. Fig. 3 shows the av- erage cluster size (ACS) for all 25 counters (grouped in 5 modules with different colors). The average of each mod- ule is indicated by the black dashed line. It decreases from about 1.2 to 1.1 depending on the applied HV which is in all cases kind of small due to the insufficient applied work- ing voltage. In addition a trend of the average cluster size is seen in each module. The MRPCs located on the top of the “hanging” module shows the largest ACS. The further down the counter is, the lower the ACS. The most prob- able reason for this observation could be the temperature gradient but also the pressure gradient in the module. Figure 3: The average cluster size of each counter in mTOF. Each point represents a counter. The gas flowed through counters from counter | to counter 25 in sequence. The dotted line is the average cluster size of five counters in each module.
Figure 4: The cluster position of all counters of mTOF. More information see text.
Figure 4: The cluster position of all counters of mTOF. More information see text.
In 2018, we have managed to produce 73 high-rate MRPC3a counters with low-resistive glass electrodes at uctech in Beijing and tested them [1-2]. They were all de- ivered to Heidelberg University and assembled into mod- ules there. As part of the CBM FAIR phase 0 programs [3], 48 of these MRPC3a counters together with 60 MRPC3b from USTC are finally installed, commissioned and oper- ated as the endcap TOF system (eTOF) at the east pole tip of the STAR experiment. The eTOF project is also part of the BES-II detector upgrade at STAR. It will pro- vide particle identification in the extended pseudorapidity range provided by the iTPC upgrade to the main tracking chamber [4]. For STAR experiment, the eTOF upgrade will bring compelling new physics capabilities to the RHIC BES-II prgram. On the other hand, CBM also benefits from STAR-eTOF by providing a large-scale integration test of the CBM-TOF hardware and software system. Figure 1: The sketch of the only one eTOF sector installed in STAR Run18.
In 2018, we have managed to produce 73 high-rate MRPC3a counters with low-resistive glass electrodes at uctech in Beijing and tested them [1-2]. They were all de- ivered to Heidelberg University and assembled into mod- ules there. As part of the CBM FAIR phase 0 programs [3], 48 of these MRPC3a counters together with 60 MRPC3b from USTC are finally installed, commissioned and oper- ated as the endcap TOF system (eTOF) at the east pole tip of the STAR experiment. The eTOF project is also part of the BES-II detector upgrade at STAR. It will pro- vide particle identification in the extended pseudorapidity range provided by the iTPC upgrade to the main tracking chamber [4]. For STAR experiment, the eTOF upgrade will bring compelling new physics capabilities to the RHIC BES-II prgram. On the other hand, CBM also benefits from STAR-eTOF by providing a large-scale integration test of the CBM-TOF hardware and software system. Figure 1: The sketch of the only one eTOF sector installed in STAR Run18.
Figure 2: (a) The XY position of the hits located at the first 2 strips of counter 101. (b) The XY position of the hits located at the last 6 strips of counter 100 while it also passes through the first 2 strips of counter 101. After the mapping from electronic address to the geome- try identifiers is applied, a series of calibrations, including local position calibration, gain correction, walk correction and TO offset correction, are carried out on the raw digis. Then the calibrated digis from different sides of the same detector strip are matched into single strip hits, and the hits are further merged into clusters. The well calibrated in- formation make it possible for us to analyze the counters’ timing performance in Run18.
Figure 2: (a) The XY position of the hits located at the first 2 strips of counter 101. (b) The XY position of the hits located at the last 6 strips of counter 100 while it also passes through the first 2 strips of counter 101. After the mapping from electronic address to the geome- try identifiers is applied, a series of calibrations, including local position calibration, gain correction, walk correction and TO offset correction, are carried out on the raw digis. Then the calibrated digis from different sides of the same detector strip are matched into single strip hits, and the hits are further merged into clusters. The well calibrated in- formation make it possible for us to analyze the counters’ timing performance in Run18.
detector hit are on the same detector, and the local distance in X and Y is smaller enough. By subtracting the eTOF arrival time and VPD start time of the match candidates the time-of-flight is got. Figure 3: The dT between eTOF and VPD as a function of pathlength. (a) before calibration. (b) after calibration.
detector hit are on the same detector, and the local distance in X and Y is smaller enough. By subtracting the eTOF arrival time and VPD start time of the match candidates the time-of-flight is got. Figure 3: The dT between eTOF and VPD as a function of pathlength. (a) before calibration. (b) after calibration.
Figure 4: The particle flight velocity as a function of its momentum. (a) before calibration. (b) after calibration.
Figure 4: The particle flight velocity as a function of its momentum. (a) before calibration. (b) after calibration.
Figure 5: The calibrated time difference distribution be- tween VPD and eTOF: (a) counter 102. (b) counter 112. (c) counter 122. Figure 6: The PID plots achieved by eTOF. (a) The match candidates’ 1/beta as a function of momentum. (b) The match candidates’ m? as a function of momentum.
Figure 5: The calibrated time difference distribution be- tween VPD and eTOF: (a) counter 102. (b) counter 112. (c) counter 122. Figure 6: The PID plots achieved by eTOF. (a) The match candidates’ 1/beta as a function of momentum. (b) The match candidates’ m? as a function of momentum.
i at pee iacesgs Hae Sies oy In general, MRPC3b is a double-stack and double-end strip readout detector. The MRPC3b mass production formally started in March 2018 after all the standards in mass pro- duction had been established. For the purpose of the quality control of the CBM-TOF MRPC, A FPGA-based TDC and DAQ system is designed by the Fast Electrical Laboratory of USTC, as illustrated in Figurel. Due to the limited time for the mass production, only 30% of the counters were measured with noise rate and took sufficient statistics for the efficiency and time resolution analysis. The batch test- ing of the MRPC3b started in April 2018. By the end of September, 80 MRPC3bs have been produced and tested and sent to Heidelberg University in several batches for module installation. Figure 1: DAQ system. Figure2 shows the test results. The anode dark currents of all tested MRPC counters are below 100nA. The reasons for the high dark current may be due to the absence of SF6 in the gas mixture and the high humidity in Hefei in sum- mer. The noise rate of all testing counters are lower than 3 Hz/cm?. The average efficiency of each counters is about 95% and the average time resolution of each counters is about 55ps. Most MRPC counters have excellent perfor- mance.
i at pee iacesgs Hae Sies oy In general, MRPC3b is a double-stack and double-end strip readout detector. The MRPC3b mass production formally started in March 2018 after all the standards in mass pro- duction had been established. For the purpose of the quality control of the CBM-TOF MRPC, A FPGA-based TDC and DAQ system is designed by the Fast Electrical Laboratory of USTC, as illustrated in Figurel. Due to the limited time for the mass production, only 30% of the counters were measured with noise rate and took sufficient statistics for the efficiency and time resolution analysis. The batch test- ing of the MRPC3b started in April 2018. By the end of September, 80 MRPC3bs have been produced and tested and sent to Heidelberg University in several batches for module installation. Figure 1: DAQ system. Figure2 shows the test results. The anode dark currents of all tested MRPC counters are below 100nA. The reasons for the high dark current may be due to the absence of SF6 in the gas mixture and the high humidity in Hefei in sum- mer. The noise rate of all testing counters are lower than 3 Hz/cm?. The average efficiency of each counters is about 95% and the average time resolution of each counters is about 55ps. Most MRPC counters have excellent perfor- mance.
Quality control in MRPC3b mass production for CBM/STAR eTOF Figure 3: MRPC voltage plateau curve.
Quality control in MRPC3b mass production for CBM/STAR eTOF Figure 3: MRPC voltage plateau curve.
'Ruprecht-Karls-Universitat Heidelberg, Germany; ?Technische Universitit Darmstadt, Germany; 3Tsinghu University, Beijing, China Figure 1: Example hit distribution over the surface of a detector after position offset calibration.
'Ruprecht-Karls-Universitat Heidelberg, Germany; ?Technische Universitit Darmstadt, Germany; 3Tsinghu University, Beijing, China Figure 1: Example hit distribution over the surface of a detector after position offset calibration.
Figure 2: corrected distribution of matched track time-of- flight - expected time of flight for pions. Ph. Weidenkaff', I. Deppner', N. Herrmann, P. Lyu’, and F. Seck*
Figure 2: corrected distribution of matched track time-of- flight - expected time of flight for pions. Ph. Weidenkaff', I. Deppner', N. Herrmann, P. Lyu’, and F. Seck*
Figure 3: system time resolution of eToF and VPD start time. With courtesy of P. Lyu.
Figure 3: system time resolution of eToF and VPD start time. With courtesy of P. Lyu.
Figure 1: 10 channels FEE board with differential outputs used in beam tests. The Front-End-Electronics (FEE) used for the MPPC readout includes the amplifier and shaper of the differen- tial signals, see photo of FEE board in Fig. 1. Due to the shaper, the signal lengh is about 0.5 jus and is about one order longer of the original signal widths after the phto- todetectors.
Figure 1: 10 channels FEE board with differential outputs used in beam tests. The Front-End-Electronics (FEE) used for the MPPC readout includes the amplifier and shaper of the differen- tial signals, see photo of FEE board in Fig. 1. Due to the shaper, the signal lengh is about 0.5 jus and is about one order longer of the original signal widths after the phto- todetectors.
The results below were obtained with first version of the readout. Figure 2: Photo of ADC64s2 board. A ToT method is an another alternative for the measure- ment of signal amplitudes. In this approach the initial pulse from MPPC is directed to shape generator. Then the shaped signal is discriminated with some threshold and time pulse is produced. This FEE is designed for work with TRB3 board.
The results below were obtained with first version of the readout. Figure 2: Photo of ADC64s2 board. A ToT method is an another alternative for the measure- ment of signal amplitudes. In this approach the initial pulse from MPPC is directed to shape generator. Then the shaped signal is discriminated with some threshold and time pulse is produced. This FEE is designed for work with TRB3 board.
Figure 3: Energy resolution of the PSD supermodule for protons as a function of beam energy. During the tests the scan of the beam momenta was done to measure the PSD supermodule response in available en- ergy range. For each beam momentum the spectrum of the deposited energies in the PSD supermodule was recorded. The mean values and the width of these spectra allow the determination of the energy resolution.
Figure 3: Energy resolution of the PSD supermodule for protons as a function of beam energy. During the tests the scan of the beam momenta was done to measure the PSD supermodule response in available en- ergy range. For each beam momentum the spectrum of the deposited energies in the PSD supermodule was recorded. The mean values and the width of these spectra allow the determination of the energy resolution.
Figure 4: Longitudinal profile of deposited energies from protons. The black and red lines correspond to proton mo- mentum 2 GeV/c and 10 GeV/c, respectively.
Figure 4: Longitudinal profile of deposited energies from protons. The black and red lines correspond to proton mo- mentum 2 GeV/c and 10 GeV/c, respectively.
It is proposed to test FEE PSD board with MPPCs dedi- cated to collect light from mPSD module and to study read- out electronics during the mCBM beam test in March 2019. The ADC PANDA ECAL board [3] will be used for MP- PCs signal readout in the beam tests in standalone mode. In order to simulate light generation without mPSD mod- ule scintillation plates will be mounted directly on MPPC diodes on FEE card. Prepared FEE with slow control box is shown in Fig. 1. View of the ADC PANDA ECAL board is presented in Fig. 2. Figure 1: FEE board with scintillators glued to MPPCs and HV control box
It is proposed to test FEE PSD board with MPPCs dedi- cated to collect light from mPSD module and to study read- out electronics during the mCBM beam test in March 2019. The ADC PANDA ECAL board [3] will be used for MP- PCs signal readout in the beam tests in standalone mode. In order to simulate light generation without mPSD mod- ule scintillation plates will be mounted directly on MPPC diodes on FEE card. Prepared FEE with slow control box is shown in Fig. 1. View of the ADC PANDA ECAL board is presented in Fig. 2. Figure 1: FEE board with scintillators glued to MPPCs and HV control box
Figure 2: ADC PANDA ECAL board
Figure 2: ADC PANDA ECAL board
Figure 3: Readout topology of ECAL PANDA ADC
Figure 3: Readout topology of ECAL PANDA ADC
TRB3 board configured to work with PANDA ECAL ADC is shown in Fig. 5. Similar readout set-u p will be used at GSI while beam test at mCBM cave in March. System will be assembled on the table to take data from cosmic muons then the detector will be installed in mC beam data. Main aim of the test is to che high readout rate and under radiation cond board needs all parts of PANDA readout BM cave to take ck the readout at ition. Now ADC chain: SODAnet and DC. As PANDA configuration based on TRBnet, clock distribution implemented in SODAnet and data transmitted via standard GTX protocol 10b8, the best way of ADC inte- gration into CBM DAQ is using GBT-FPGA module what applied in CBM readout system. Main points of ADC inte- gration are: GBT integration in ADC FPGA project using on board clock, DAQ clock distribution to figuration procedure. This work will be beam at March 19. ACD, ADC con- started after test Figure 4: Readout ADC workbench schematic
TRB3 board configured to work with PANDA ECAL ADC is shown in Fig. 5. Similar readout set-u p will be used at GSI while beam test at mCBM cave in March. System will be assembled on the table to take data from cosmic muons then the detector will be installed in mC beam data. Main aim of the test is to che high readout rate and under radiation cond board needs all parts of PANDA readout BM cave to take ck the readout at ition. Now ADC chain: SODAnet and DC. As PANDA configuration based on TRBnet, clock distribution implemented in SODAnet and data transmitted via standard GTX protocol 10b8, the best way of ADC inte- gration into CBM DAQ is using GBT-FPGA module what applied in CBM readout system. Main points of ADC inte- gration are: GBT integration in ADC FPGA project using on board clock, DAQ clock distribution to figuration procedure. This work will be beam at March 19. ACD, ADC con- started after test Figure 4: Readout ADC workbench schematic
Figure 5: TRB3 based DAQ for PANDA ECAL ADC
Figure 5: TRB3 based DAQ for PANDA ECAL ADC
The Prony method has no information about the start time of the signal, so at the first iteration the start time is determined by the linear spline of the leading edge of the signal. After calculating the attenuation coefficients and the corresponding amplitudes, good events are selected by applying the Pearson criterion. Correctly determined atten- uation coefficients are entered into the ROOT histograms. The second iteration of signal processing consists in the im- plementation of the third stage of the Prony method, with an indication of the average values of the attenuation co- efficients determined at the first iteration. The start time of the signal to is determined by the maximum signal time and the known attenuation coefficients: Figure 1: An example of waveform (black) and superim- posed fit (red). Figure | shows the characteristic waveform and super- imposed fit. The obtained chi square values, normalized to the number of degrees of freedom, are shown in Figure 2, depending on the signal area under the fit. One can see a good separation of the muon spot located near the 3000 ADC channel and noise events grouped around 0. At the
The Prony method has no information about the start time of the signal, so at the first iteration the start time is determined by the linear spline of the leading edge of the signal. After calculating the attenuation coefficients and the corresponding amplitudes, good events are selected by applying the Pearson criterion. Correctly determined atten- uation coefficients are entered into the ROOT histograms. The second iteration of signal processing consists in the im- plementation of the third stage of the Prony method, with an indication of the average values of the attenuation co- efficients determined at the first iteration. The start time of the signal to is determined by the maximum signal time and the known attenuation coefficients: Figure 1: An example of waveform (black) and superim- posed fit (red). Figure | shows the characteristic waveform and super- imposed fit. The obtained chi square values, normalized to the number of degrees of freedom, are shown in Figure 2, depending on the signal area under the fit. One can see a good separation of the muon spot located near the 3000 ADC channel and noise events grouped around 0. At the
same time, the chi square selection also makes it possible to reliably separate these two groups of events. Figure 2: Normalized chi square as a function of signal charge.
same time, the chi square selection also makes it possible to reliably separate these two groups of events. Figure 2: Normalized chi square as a function of signal charge.
Figure 2: PSD Layouts (7x9 modules layout on left side, 6x8 modules layout on right side). Figure 1: PSD Platform Design Study Overview.
Figure 2: PSD Layouts (7x9 modules layout on left side, 6x8 modules layout on right side). Figure 1: PSD Platform Design Study Overview.
The PSD platform (see design study overview in Fig. 1) consists of three independent frames for movement of PSD in X and Y and Y rotation. Z movement is realized on the support structure the PSD platform is sitting on and it is not part of the PSD platform design. The PSD platform support structure design is going to be realized as soon as a CBM cave rails system is defined.
The PSD platform (see design study overview in Fig. 1) consists of three independent frames for movement of PSD in X and Y and Y rotation. Z movement is realized on the support structure the PSD platform is sitting on and it is not part of the PSD platform design. The PSD platform support structure design is going to be realized as soon as a CBM cave rails system is defined.
Figure 3: Reachable positions of PSD platform (all pictures show 7x9 modules PSD layout). The PSD platform 3D model is shown in Fig. 3. The developed platform allows to move with PSD in ranges +/- 900mm along X axis, +/- 700mm along Y axis and to ro- tate with PSD in range +/- 3deg along Y axis. The cage (in yellow) is able to accommodate two different PSD layouts as requested and allows to move with it by crane indepen- dently on the rest of the platform. Total weight of PSD and platform is approx. 29tons (23tons PSD + 6tons the platform).
Figure 3: Reachable positions of PSD platform (all pictures show 7x9 modules PSD layout). The PSD platform 3D model is shown in Fig. 3. The developed platform allows to move with PSD in ranges +/- 900mm along X axis, +/- 700mm along Y axis and to ro- tate with PSD in range +/- 3deg along Y axis. The cage (in yellow) is able to accommodate two different PSD layouts as requested and allows to move with it by crane indepen- dently on the rest of the platform. Total weight of PSD and platform is approx. 29tons (23tons PSD + 6tons the platform).
Figure 1: Distribution of ionizing energy loss in transverse plane at 20 cm depth from the PSD module front. The calorimeter irradiation was simulated with the FLUKA code for 2 months of the CBM run with Au+Au collisions on a | % interaction Au target at beam energy of 10 AGeV at the beam rate of 10° ions/s. Fig. 1 shows that the absorbed dose in the PSD does not exceed 1 kGy. Dam- age by the ionizing dose which is the most crucial for the scintillator tiles was studied in details for calorimeters of LHCb experiment [2]. The relative light yield of the LHCb tiles degraded by 25 % after 2.5 kGy irradiation, and then slowly decreased further by 20 % after irradiation up to 14 kGy. Therefore, scintillators and WLS-fibers of PSD are not expected to degrade significantly. Moreover, the modu- lar structure and the longitudinal segmentation of PSD cou- pled with the regular calibration will ensure the stability of transverse and longitudinal uniformity of the light col- lection. Scintillator irradiation tests are being discussed to study their degradation in details.
Figure 1: Distribution of ionizing energy loss in transverse plane at 20 cm depth from the PSD module front. The calorimeter irradiation was simulated with the FLUKA code for 2 months of the CBM run with Au+Au collisions on a | % interaction Au target at beam energy of 10 AGeV at the beam rate of 10° ions/s. Fig. 1 shows that the absorbed dose in the PSD does not exceed 1 kGy. Dam- age by the ionizing dose which is the most crucial for the scintillator tiles was studied in details for calorimeters of LHCb experiment [2]. The relative light yield of the LHCb tiles degraded by 25 % after 2.5 kGy irradiation, and then slowly decreased further by 20 % after irradiation up to 14 kGy. Therefore, scintillators and WLS-fibers of PSD are not expected to degrade significantly. Moreover, the modu- lar structure and the longitudinal segmentation of PSD cou- pled with the regular calibration will ensure the stability of transverse and longitudinal uniformity of the light col- lection. Scintillator irradiation tests are being discussed to study their degradation in details.
Figure 2: Non-ionizing energy loss vs radius at transverse plane where MPPCs are located for different configurations of the beam hole. To avoid a high radiation load in the central modules, there is a beam hole in the center of PSD. Three different beam hole shapes were considered, namely round with a 3 cm radius, quadratic- and diamond-shaped of 20 x 20 cm? size. Fig. 2 illustrates the reduction of non-ionizing energy loss by factor of 2 at the distance of 10 cm from the beam axis for the enlarged beam hole. Diamond-shaped hole de- sign was chosen because it provided additional reduction to the ionizing dose compared to the quadratic-shaped de- sign [3]. Another safety measure in form of a dedicated neutron shielding based on 8 cm thick polyethylene blocks with 3 % of boron is introduced reducing the fluence up to factor of five.
Figure 2: Non-ionizing energy loss vs radius at transverse plane where MPPCs are located for different configurations of the beam hole. To avoid a high radiation load in the central modules, there is a beam hole in the center of PSD. Three different beam hole shapes were considered, namely round with a 3 cm radius, quadratic- and diamond-shaped of 20 x 20 cm? size. Fig. 2 illustrates the reduction of non-ionizing energy loss by factor of 2 at the distance of 10 cm from the beam axis for the enlarged beam hole. Diamond-shaped hole de- sign was chosen because it provided additional reduction to the ionizing dose compared to the quadratic-shaped de- sign [3]. Another safety measure in form of a dedicated neutron shielding based on 8 cm thick polyethylene blocks with 3 % of boron is introduced reducing the fluence up to factor of five.
wide range of 6 x 10/°-9 x 10!? n.,/cm?. After irradi- ation with 2 x 10" n.4/cm? MPPC dark current increased by 3 orders of magnitude and reached 10 A [4]. This re- sulted in the increase of noise by 1 order of magnitude and consequent drop of signal to noise ratio down to ~ 50. Figure 3: Single module proton energy resolution vs beam energies for different neutron fluences acquired by MPPCs (top) and vs fluence for different beam energies (bottom).
wide range of 6 x 10/°-9 x 10!? n.,/cm?. After irradi- ation with 2 x 10" n.4/cm? MPPC dark current increased by 3 orders of magnitude and reached 10 A [4]. This re- sulted in the increase of noise by 1 order of magnitude and consequent drop of signal to noise ratio down to ~ 50. Figure 3: Single module proton energy resolution vs beam energies for different neutron fluences acquired by MPPCs (top) and vs fluence for different beam energies (bottom).
Figure 1: Schematic view of the BM@N set up. Schematic view of the BM@N experimental setup is shown in Fig. 1. The momentum measurements of pro- duced charged particles is performed with high preci- sion track measurements by central tracker based on two- coordinate triple GEM detectors located downstream of the target inside the magnet and the drift/cathode pad chambers (DCH/CPC) of the outer tracker placed behind the magnet. BM@N dipole magnet has the gap around 1 m between the poles. The maximum magnetic field of the magnet is up to 1.2 T and can be varied to get the optimal BM@N detector acceptance and momentum resolution for differ- ent processes and beam energies. Particle identification is done by TOF measured between TO detectors and multi- gap resistive plate chambers (mRPC) with a strip read-out. The TOF detectors provide time resolution about 80 psec. and allow to discriminate hadrons (7, K, p) as well as light nuclei with the momentum up to few GeV/c produced in multi-particle events. To measure the collision centrality in nucleus-nucleus collisions the ZDC is used. This hadron calorimeter is placed at the end of BM@N beam line on distance 9 m from the target.
Figure 1: Schematic view of the BM@N set up. Schematic view of the BM@N experimental setup is shown in Fig. 1. The momentum measurements of pro- duced charged particles is performed with high preci- sion track measurements by central tracker based on two- coordinate triple GEM detectors located downstream of the target inside the magnet and the drift/cathode pad chambers (DCH/CPC) of the outer tracker placed behind the magnet. BM@N dipole magnet has the gap around 1 m between the poles. The maximum magnetic field of the magnet is up to 1.2 T and can be varied to get the optimal BM@N detector acceptance and momentum resolution for differ- ent processes and beam energies. Particle identification is done by TOF measured between TO detectors and multi- gap resistive plate chambers (mRPC) with a strip read-out. The TOF detectors provide time resolution about 80 psec. and allow to discriminate hadrons (7, K, p) as well as light nuclei with the momentum up to few GeV/c produced in multi-particle events. To measure the collision centrality in nucleus-nucleus collisions the ZDC is used. This hadron calorimeter is placed at the end of BM@N beam line on distance 9 m from the target.
Starting from 2020 new heavy ion experiments at BM@N are planned. It was shown by FLUKA simulations, that the radiation doses from ionizing and non-ionizing par- ticles in central part of the present ZDC expected for gold beam at energy 4 AGeV and beam rate 2x10° ions per sec- ond will be too high and the light output in central modules will be decreased due to radiation damage.
Starting from 2020 new heavy ion experiments at BM@N are planned. It was shown by FLUKA simulations, that the radiation doses from ionizing and non-ionizing par- ticles in central part of the present ZDC expected for gold beam at energy 4 AGeV and beam rate 2x10° ions per sec- ond will be too high and the light output in central modules will be decreased due to radiation damage.
Figure 3: Photo of PSD module installed at BM@N beam line.
Figure 3: Photo of PSD module installed at BM@N beam line.
Figure 5: The longitudinal shower profile in PSD module for 3.5 AGeV *° Ar ions.
Figure 5: The longitudinal shower profile in PSD module for 3.5 AGeV *° Ar ions.
Within the CBM Phase 0 program various modules of the CBM Project Spectator Detector (PSD) will be used in the several experiments: in the fixed-target Baryonic Mat- ter experiment at the Nuclotron (BM@N), in the Heavy Ion and Neutrino Experiment NA61 (Shine) at SPS, and in the mCBM experiment at SIS18. The geometries of the differ- ent PSD versions are shown in figure | for the BM@N ex- periment (top, CBM modules in yellow), and for the NA61 experiment(bottom, CBM modules in red). The CBM PSD module used in the mCBM setup at SIS18 is shown figure 2 together with the other detectors. Figure 1: The geometries of the different PSD versions in BME@N (top, CBM modules in yellow) and NA61 (bottom, CBM modules in red).
Within the CBM Phase 0 program various modules of the CBM Project Spectator Detector (PSD) will be used in the several experiments: in the fixed-target Baryonic Mat- ter experiment at the Nuclotron (BM@N), in the Heavy Ion and Neutrino Experiment NA61 (Shine) at SPS, and in the mCBM experiment at SIS18. The geometries of the differ- ent PSD versions are shown in figure | for the BM@N ex- periment (top, CBM modules in yellow), and for the NA61 experiment(bottom, CBM modules in red). The CBM PSD module used in the mCBM setup at SIS18 is shown figure 2 together with the other detectors. Figure 1: The geometries of the different PSD versions in BME@N (top, CBM modules in yellow) and NA61 (bottom, CBM modules in red).
Figure 2: The mCBM setup with the PSD module (GEANT3 picture). Figure 3: The geometry of the PSD as implemented in FLUKA.
Figure 2: The mCBM setup with the PSD module (GEANT3 picture). Figure 3: The geometry of the PSD as implemented in FLUKA.
GSI, Darmstadt, Germany
GSI, Darmstadt, Germany
The neutron equivalent dose distribution in the place of photon-counters is shown in figure 4 for the various exper- imental conditions discussed above. Tests of counters have demonstrated that the internal noise increases after irradi- ation with doses above 101! Neg/em?. It is worthwhile to note, that the non-ionizing dose for the modules to be used in the BM@N experiment differ only about a factor of 2 for the detector configurations with hole (v1) and without (v1), whereas for the ionizing dose the difference was a factor of 100. Figure 4: Non-ionizing dose in MPPC of CBM PSD in FAIR PhaseO experiments.
The neutron equivalent dose distribution in the place of photon-counters is shown in figure 4 for the various exper- imental conditions discussed above. Tests of counters have demonstrated that the internal noise increases after irradi- ation with doses above 101! Neg/em?. It is worthwhile to note, that the non-ionizing dose for the modules to be used in the BM@N experiment differ only about a factor of 2 for the detector configurations with hole (v1) and without (v1), whereas for the ionizing dose the difference was a factor of 100. Figure 4: Non-ionizing dose in MPPC of CBM PSD in FAIR PhaseO experiments.
a [he magnet has a free aperture of 1.44 m vertically and 3.0 m horizontally in order to accommodate the STS de- tector system with a polar angle acceptance of 25 degrees and a horizontal acceptance of 30 degrees. The 3D mag- netic field calculations were made with two different codes (Mermaid and ANSYS), whereas the forces on the coils and the poles were calculated with the ANSYS 3D model. The results of the two codes for the magnetic field distribu- tion along the beam are shown in figure 1. The maximum field in the centre of the magnet is 1.08 T (Mermaid) and 1.12 T (ANSYS). Figure 1: Magnetic field distribution along the beam start- ing in the center of the magnet. Blue line: Mermaid calcu- lation. Red line: ANSYS calculation.
a [he magnet has a free aperture of 1.44 m vertically and 3.0 m horizontally in order to accommodate the STS de- tector system with a polar angle acceptance of 25 degrees and a horizontal acceptance of 30 degrees. The 3D mag- netic field calculations were made with two different codes (Mermaid and ANSYS), whereas the forces on the coils and the poles were calculated with the ANSYS 3D model. The results of the two codes for the magnetic field distribu- tion along the beam are shown in figure 1. The maximum field in the centre of the magnet is 1.08 T (Mermaid) and 1.12 T (ANSYS). Figure 1: Magnetic field distribution along the beam start- ing in the center of the magnet. Blue line: Mermaid calcu- lation. Red line: ANSYS calculation.
Figure 2: General view of the CBM magnet including cryo- genics provided by BIBP.
Figure 2: General view of the CBM magnet including cryo- genics provided by BIBP.
The Conceptual Design Report has been approved in April 2018. The manufacture of the superconducting cable and of the iron yoke has started. The magnet will be pro- vided by the Budker-Institute for Nuclear Physics (BINP) in Novosibirsk. Figure 3: The superconducting upper coil with supports struts.
The Conceptual Design Report has been approved in April 2018. The manufacture of the superconducting cable and of the iron yoke has started. The magnet will be pro- vided by the Budker-Institute for Nuclear Physics (BINP) in Novosibirsk. Figure 3: The superconducting upper coil with supports struts.
Figure 1: 3D model of CBM dipole magnet. where J(Z) is the current density.
Figure 1: 3D model of CBM dipole magnet. where J(Z) is the current density.
The magnet field saturation inside the CBM magnet is shown in Fig.2. Maximal field saturation is at the poles. Figure 3: Field distribution inside the CBM dipole magnet obtained by VIEM [4].
The magnet field saturation inside the CBM magnet is shown in Fig.2. Maximal field saturation is at the poles. Figure 3: Field distribution inside the CBM dipole magnet obtained by VIEM [4].
of the magnet has been divided into 12841 tetrahedrons. The total number of vertexes was 3128. 3D model of CBM dipole magnet is presented in Fig. 1. Figure 2: The CBM magnet saturation picture.
of the magnet has been divided into 12841 tetrahedrons. The total number of vertexes was 3128. 3D model of CBM dipole magnet is presented in Fig. 1. Figure 2: The CBM magnet saturation picture.
Figure 4: The comparison of the results obtained by VIEM [4] and with TOSCA [3] for the vertical component of the magnetic field.
Figure 4: The comparison of the results obtained by VIEM [4] and with TOSCA [3] for the vertical component of the magnetic field.
Figure 2: The field distribution in the area of RICH detec- tor. Figure 1: 3D model of CBM dipole magnet with RICH photodetectors.
Figure 2: The field distribution in the area of RICH detec- tor. Figure 1: 3D model of CBM dipole magnet with RICH photodetectors.
The CBM superconducting dipole magnet has to provide the bending power of ~1 T-m and quite low stray magnetic field in the region of RICH [1]. The magnetic field in the vicinity of the RICH photodetector has been calculated for BINP design of the CBM magnet. The 3D- model of CBM dipole magnet with RICH photodetector without z- sym- metry for the TOSCA [2] simulation is presented in Fig: 1.
The CBM superconducting dipole magnet has to provide the bending power of ~1 T-m and quite low stray magnetic field in the region of RICH [1]. The magnetic field in the vicinity of the RICH photodetector has been calculated for BINP design of the CBM magnet. The 3D- model of CBM dipole magnet with RICH photodetector without z- sym- metry for the TOSCA [2] simulation is presented in Fig: 1.
Figure 5: View of the magnetic field distribution in the area of the RICH shielding box from the side. Fig.4 shows the distribution of the absolute value of magnetic field inside the shielding box for the BINP ver-
Figure 5: View of the magnetic field distribution in the area of the RICH shielding box from the side. Fig.4 shows the distribution of the absolute value of magnetic field inside the shielding box for the BINP ver-
Figure 4: Distribution of the magnetic field in the area of the RICH shielding box. Figure 3: The shielding box of the RICH photodetector.
Figure 4: Distribution of the magnetic field in the area of the RICH shielding box. Figure 3: The shielding box of the RICH photodetector.
Figure 6: The distribution of B, magnetic field component along z axis for the electron version of the CBM detector. Z position is given in mm.
Figure 6: The distribution of B, magnetic field component along z axis for the electron version of the CBM detector. Z position is given in mm.
Figure 2: Total magnetic flux density in the CBM magnet. Figure 1: BINP design of the CBM magnet.
Figure 2: Total magnetic flux density in the CBM magnet. Figure 1: BINP design of the CBM magnet.
BINP design of the CBM superconducting dipole mag- net with the enlarged vertical aperture up to 1440 mm is shown in Fig.1. It has few differences from the initial one [1], namely, the cylindrical poles, 4 flat vertical beams and 6 narrow horizontal beams. The materials of the poles and yoke are the ARMCO and the SA1010 steel, respectively. Magnetic and stress analysis for BINP design was made with the ANSYS [2] and TOSCA [3].
BINP design of the CBM superconducting dipole mag- net with the enlarged vertical aperture up to 1440 mm is shown in Fig.1. It has few differences from the initial one [1], namely, the cylindrical poles, 4 flat vertical beams and 6 narrow horizontal beams. The materials of the poles and yoke are the ARMCO and the SA1010 steel, respectively. Magnetic and stress analysis for BINP design was made with the ANSYS [2] and TOSCA [3].
Figure 3: Total magnetic flux density in the CBM magnet coil.
Figure 3: Total magnetic flux density in the CBM magnet coil.
Figure 4: The CBM magnet saturation picture.
Figure 4: The CBM magnet saturation picture.
Figure 5: The vertical component of the magnetic field dis- tribution for the muon version of the CBM detector. z po- sition is given in mm.
Figure 5: The vertical component of the magnetic field dis- tribution for the muon version of the CBM detector. z po- sition is given in mm.
The CBM beam dump was designed for high-intensity heavy-ion beams from SIS300, i.e. Au beams with a kinetic energy of 30 A GeV and an intensity of 10° ions/s. Cross sections of the CBM cave and the beam dump are shown in figure 1. The dump consists of an iron core (4000 tons) surrounded by a concrete shielding (5000 tons). The total length of the dump is approximately 30 m with a width and height of about 10 m, the entrance window has a size of 2 x2m?. Figure 1: Side view (upper panel) and top view (lower panel) of the CBM cave and the beam dump (see text).
The CBM beam dump was designed for high-intensity heavy-ion beams from SIS300, i.e. Au beams with a kinetic energy of 30 A GeV and an intensity of 10° ions/s. Cross sections of the CBM cave and the beam dump are shown in figure 1. The dump consists of an iron core (4000 tons) surrounded by a concrete shielding (5000 tons). The total length of the dump is approximately 30 m with a width and height of about 10 m, the entrance window has a size of 2 x2m?. Figure 1: Side view (upper panel) and top view (lower panel) of the CBM cave and the beam dump (see text).
Figure 2: Horizontal distribution of the non-ionizing dose in the CBM cave without shielding (see text). and its electronics will be irradiated with a total dose be- tween about 5x 1010 and 5x 1011 neq/cm?. It is important to note, that in this calculation the vertical opening of the beam dump already has been reduced to a height of 20 cm only by concrete blocks of | m thickness, as illustrated in figure 4.
Figure 2: Horizontal distribution of the non-ionizing dose in the CBM cave without shielding (see text). and its electronics will be irradiated with a total dose be- tween about 5x 1010 and 5x 1011 neq/cm?. It is important to note, that in this calculation the vertical opening of the beam dump already has been reduced to a height of 20 cm only by concrete blocks of | m thickness, as illustrated in figure 4.
Figure 3: Horizontal distribution of the dose rate in the CBM cave due to the activation of the beam dump and the target 1 day after beam shut down (see text). Figure 3 illustrates the horizontal distribution of dose rate due to the activation of the iron beam dump and the tar- get 1 day after beam shut down. The FLUKA calculation has been performed for an Au beam with kinetic energy of 2A GeV and an intensity of 10° ions/s over 2 months for a 1% interaction Au target and 50% of the nominal magnetic field. The dose rate downstream the TOF wall increases from about 5 Sv/h up to about 50 wSv/h close to the beam dump entrance.
Figure 3: Horizontal distribution of the dose rate in the CBM cave due to the activation of the beam dump and the target 1 day after beam shut down (see text). Figure 3 illustrates the horizontal distribution of dose rate due to the activation of the iron beam dump and the tar- get 1 day after beam shut down. The FLUKA calculation has been performed for an Au beam with kinetic energy of 2A GeV and an intensity of 10° ions/s over 2 months for a 1% interaction Au target and 50% of the nominal magnetic field. The dose rate downstream the TOF wall increases from about 5 Sv/h up to about 50 wSv/h close to the beam dump entrance.
pipe. The horizontal width of the slit is required for mov- ing the beam pipe according to the beam deflection. After beam shut down, the slit will be completely closed by two movable paraffin wax blocks. During the experiment, the paraffin blocks will be shifted in order to open space for the beam pipe. The paraffin shielding will reduce the radiation level due to backscattered particles during beam operation. Figure 4: Front view of the beam dump window with con- crete and paraffin shielding during beam operation. Please note that the figure is not to scale. The concrete blocks cover the 2 x 2 m? large opening of the beam dump, except for the window for the beam, which has a width of 2 m and a height of 20 cm. After beam shut down, the hole for the beam will be closed by the mobile paraffin blocks (see text and figure description).
pipe. The horizontal width of the slit is required for mov- ing the beam pipe according to the beam deflection. After beam shut down, the slit will be completely closed by two movable paraffin wax blocks. During the experiment, the paraffin blocks will be shifted in order to open space for the beam pipe. The paraffin shielding will reduce the radiation level due to backscattered particles during beam operation. Figure 4: Front view of the beam dump window with con- crete and paraffin shielding during beam operation. Please note that the figure is not to scale. The concrete blocks cover the 2 x 2 m? large opening of the beam dump, except for the window for the beam, which has a width of 2 m and a height of 20 cm. After beam shut down, the hole for the beam will be closed by the mobile paraffin blocks (see text and figure description).
In order to estimate the required thickness of the paraffin blocks, FLUKA calculations have been performed for an Au beam with a kinetic energy of 11A GeV, and a beam intensity of 10° ions/s integrated over a beam time of 2 months. Figure 5 depicts the dose rate in front of the shield- ing 1 day after beam shut down for different thicknesses of the paraffin blocks. Assuming a thickness of 60 cm for the paraffin shielding, the dose can be reduced by a factor of about 100 down to a value below 3 jSv/h in front of the paraffin blocks. Figure 5: Distribution of the dose rate in the CBM cave due to the activation of the beam dump for different thicknesses of the paraffin shielding calculated with FLUKA 1 day after beam shut down (see text).
In order to estimate the required thickness of the paraffin blocks, FLUKA calculations have been performed for an Au beam with a kinetic energy of 11A GeV, and a beam intensity of 10° ions/s integrated over a beam time of 2 months. Figure 5 depicts the dose rate in front of the shield- ing 1 day after beam shut down for different thicknesses of the paraffin blocks. Assuming a thickness of 60 cm for the paraffin shielding, the dose can be reduced by a factor of about 100 down to a value below 3 jSv/h in front of the paraffin blocks. Figure 5: Distribution of the dose rate in the CBM cave due to the activation of the beam dump for different thicknesses of the paraffin shielding calculated with FLUKA 1 day after beam shut down (see text).
without paraffin beam dump shielding. The FLUKA calcu- lation was performed for an Au beam with kinetic energy of 2A GeV and an intensity of 10° ions/s. The upper and lower panel of figure 6 depicts the rate density of charged particles and neutrons, respectively. The profiles of the particle distributions indicate that the charged particles are emitted predominantly from the target region, whereas the neutrons stem from the beam dump. This scenario is sup- ported by figure 7 which presents the same distributions as in figure 6 but with paraffin beam dump shielding of 60 cm thickness. The charged particle density in the cave is not affected by the beam dump shielding, whereas the neutron radiation is reduced by shielding the dump. Figure 6: Rate density of charged particles (upper panel) and neutrons (lower panel) in the CBM cave during the measurements without paraffin beam dump shielding (see text).
without paraffin beam dump shielding. The FLUKA calcu- lation was performed for an Au beam with kinetic energy of 2A GeV and an intensity of 10° ions/s. The upper and lower panel of figure 6 depicts the rate density of charged particles and neutrons, respectively. The profiles of the particle distributions indicate that the charged particles are emitted predominantly from the target region, whereas the neutrons stem from the beam dump. This scenario is sup- ported by figure 7 which presents the same distributions as in figure 6 but with paraffin beam dump shielding of 60 cm thickness. The charged particle density in the cave is not affected by the beam dump shielding, whereas the neutron radiation is reduced by shielding the dump. Figure 6: Rate density of charged particles (upper panel) and neutrons (lower panel) in the CBM cave during the measurements without paraffin beam dump shielding (see text).
Figure 7: Rate density of charged particles (upper panel) and neutrons (lower panel) in the CBM cave during the measurements with a 60 cm thick paraffin beam dump shielding (see text).
Figure 7: Rate density of charged particles (upper panel) and neutrons (lower panel) in the CBM cave during the measurements with a 60 cm thick paraffin beam dump shielding (see text).
Figure 1: Horizontal profile of an Au beam with kinetic energy 2A GeV at a distance of 1.6 m downstream a 250 jum thick Au target for full magnetic field calculated with the FLUKA code assuming a typical SIS100 Au beam with Ax = Ay = 0.6 mm and a divergence of 1.7 mrad (blue histogram) compared to a point-like beam (red histogram).
Figure 1: Horizontal profile of an Au beam with kinetic energy 2A GeV at a distance of 1.6 m downstream a 250 jum thick Au target for full magnetic field calculated with the FLUKA code assuming a typical SIS100 Au beam with Ax = Ay = 0.6 mm and a divergence of 1.7 mrad (blue histogram) compared to a point-like beam (red histogram).
Figure 2: Left: Deflection of Au beams with kinetic ener- gies between 2A and 8A GeV by the magnetic field with an integral of BL = 1 Tm (= 100%). Right: Horizontal profiles of Au beams with kinetic energies from 2A to 10A GeV at the entrance of the beam dump traversing a 250 jum hick Au target for different scaling factors of the magnetic field as proposed in the TDR of the PSD. The distributions are normalized to one Au ion.
Figure 2: Left: Deflection of Au beams with kinetic ener- gies between 2A and 8A GeV by the magnetic field with an integral of BL = 1 Tm (= 100%). Right: Horizontal profiles of Au beams with kinetic energies from 2A to 10A GeV at the entrance of the beam dump traversing a 250 jum hick Au target for different scaling factors of the magnetic field as proposed in the TDR of the PSD. The distributions are normalized to one Au ion.
is followed by the beam pipe within the RICH or the MuCh detector. These two beam pipes have to fulfil very different requirements concerning material budget, and will be de- signed for each detector separately. However, these beam pipes should agree in their general shape in order to fit to the upstream STS beam pipe and to the downstream TRD and TOF beam pipe. For the following design considera- tions and corresponding calculations it is assumed that the RICH/MuCh beam pipe continues like the conical part of the STS beam pipe, resulting in a radius of R = 7.47 cm at a distance of 170 cm downstream the target, and a radius of R = 16.16 cm at a distance of 370 cm. The beam profile at 370 cm is shown in figure 4. For the calculations it is assumed that the beam pipe consists of carbon fiber with 0.5 mm wall thickness. Figure 4: Au beam profiles for kinetic energies between 2A and 10 A GeV at a distance of 370 cm behind the 250 wm thick Au target for scaled magnetic fields. The distributions are normalized to one Au ion.
is followed by the beam pipe within the RICH or the MuCh detector. These two beam pipes have to fulfil very different requirements concerning material budget, and will be de- signed for each detector separately. However, these beam pipes should agree in their general shape in order to fit to the upstream STS beam pipe and to the downstream TRD and TOF beam pipe. For the following design considera- tions and corresponding calculations it is assumed that the RICH/MuCh beam pipe continues like the conical part of the STS beam pipe, resulting in a radius of R = 7.47 cm at a distance of 170 cm downstream the target, and a radius of R = 16.16 cm at a distance of 370 cm. The beam profile at 370 cm is shown in figure 4. For the calculations it is assumed that the beam pipe consists of carbon fiber with 0.5 mm wall thickness. Figure 4: Au beam profiles for kinetic energies between 2A and 10 A GeV at a distance of 370 cm behind the 250 wm thick Au target for scaled magnetic fields. The distributions are normalized to one Au ion.
Figure 3: Beam pipe within the STS.
Figure 3: Beam pipe within the STS.
to hit the beam pipe. Moreover, when continuing with a conical beam pipe, it would not fit into the inner holes of TRD, TOF and PSD. Therefore, it is assumed for the fol- lowing simulations that the beam pipe will continue with a cylindrical shape, which is tilted to confine the beam in the center until it reaches the beam dump. Figure 5 depicts two beam pipe options with a bellow between RICH/MuCh and the TRD detector. One concept is a cylindrical beam pipe with a radius of R = 16.16 cm up to the entrance of the PSD, where the radius is reduced to R = 9.5 cm in order to fit into the hole of the PSD. The reduction of the radius is accompanied by a shift of the beam pipe in order to keep the beam in the center of the smaller pipe. The other con- cept assumes the reduction of the radius to R = 9.5 cm and the re-centering of the pipe directly behind the bellow. For both beam pipe options the bellow has to provide a kink angle between 0.7 and 1.8 degrees for Au beams with en- ergies of 1OA and 2A GeV, respectively. The tilt of the beam pipe according to the beam rigidity requires the cor- responding movement of the mechanical structures of the pipe including pumps, and of the PSD. For the following simulations it is assumed that the beam pipes consists of carbon fiber with 0.5 mm wall thickness. Figure 5: Models of the two beam pipe versions in CBM- root.
to hit the beam pipe. Moreover, when continuing with a conical beam pipe, it would not fit into the inner holes of TRD, TOF and PSD. Therefore, it is assumed for the fol- lowing simulations that the beam pipe will continue with a cylindrical shape, which is tilted to confine the beam in the center until it reaches the beam dump. Figure 5 depicts two beam pipe options with a bellow between RICH/MuCh and the TRD detector. One concept is a cylindrical beam pipe with a radius of R = 16.16 cm up to the entrance of the PSD, where the radius is reduced to R = 9.5 cm in order to fit into the hole of the PSD. The reduction of the radius is accompanied by a shift of the beam pipe in order to keep the beam in the center of the smaller pipe. The other con- cept assumes the reduction of the radius to R = 9.5 cm and the re-centering of the pipe directly behind the bellow. For both beam pipe options the bellow has to provide a kink angle between 0.7 and 1.8 degrees for Au beams with en- ergies of 1OA and 2A GeV, respectively. The tilt of the beam pipe according to the beam rigidity requires the cor- responding movement of the mechanical structures of the pipe including pumps, and of the PSD. For the following simulations it is assumed that the beam pipes consists of carbon fiber with 0.5 mm wall thickness. Figure 5: Models of the two beam pipe versions in CBM- root.
sibilities to run the GUI: the user has to copy all the files needed into his home directory, and run GUI without con- nection to GSI. Alternatively, the user can run the macro after connection to GSI on the LINUX server. The full paths to the files are presented in the note. The ROOT inter- face provides more possibilities as the Web interface (see below). Figure 3 shows the GUI after running the macro. The window comprises 2 parts: input parameters on the left side, and FLUKA results on the right side. Figure 2: GUI after run ROOT macro.
sibilities to run the GUI: the user has to copy all the files needed into his home directory, and run GUI without con- nection to GSI. Alternatively, the user can run the macro after connection to GSI on the LINUX server. The full paths to the files are presented in the note. The ROOT inter- face provides more possibilities as the Web interface (see below). Figure 3 shows the GUI after running the macro. The window comprises 2 parts: input parameters on the left side, and FLUKA results on the right side. Figure 2: GUI after run ROOT macro.
Figure 1: Web interface for the FLUKA results.
Figure 1: Web interface for the FLUKA results.
Figure 3: GUI with geometry of selected area and 2D his- togram for selected Z position. The user can select with a mouse click the position of the equipment from the geometry picture: Z position in case of cave and magnet gap, and Y position for E10 service area. The corresponding coordinate will be updated in the “Set position...” group. After pressing the “Draw” button, the 2D histogram for the selected region will be presented in the window. With a mouse click the user can select a bin of interests, and the value from this bin will be shown in the “FLUKA results” window with errors. An example is shown in figure 3.
Figure 3: GUI with geometry of selected area and 2D his- togram for selected Z position. The user can select with a mouse click the position of the equipment from the geometry picture: Z position in case of cave and magnet gap, and Y position for E10 service area. The corresponding coordinate will be updated in the “Set position...” group. After pressing the “Draw” button, the 2D histogram for the selected region will be presented in the window. With a mouse click the user can select a bin of interests, and the value from this bin will be shown in the “FLUKA results” window with errors. An example is shown in figure 3.
Figure 4: The same window, as in Figure 3, but with setting of safety level for NIEL. The user can set the safety limit for equipment in “Set dose value” window. After pressing the “Draw” button the 2D histogram with values above limit will be presented in window (see figure 4).
Figure 4: The same window, as in Figure 3, but with setting of safety level for NIEL. The user can set the safety limit for equipment in “Set dose value” window. After pressing the “Draw” button the 2D histogram with values above limit will be presented in window (see figure 4).
the The 3D CAD view (figure 1) shows the area below the M service platform and the neighbouring parts of the M setup, especially magnet foundation, CBM dipole CBM RICH and MUCH detectors. Figure 2 shows anticipated locations of service electronics racks, the two main rack rows are also indicated in the CAD view (in dar! k violet). Based in the CAD model a FLUKA geometry was created, only containing the geometry and the materi- als relevant for radiation calculations. Figure 3 shows the Z-X and the Z-Y views. Figure 1: 3D CAD view of the space below CBM service platform.
the The 3D CAD view (figure 1) shows the area below the M service platform and the neighbouring parts of the M setup, especially magnet foundation, CBM dipole CBM RICH and MUCH detectors. Figure 2 shows anticipated locations of service electronics racks, the two main rack rows are also indicated in the CAD view (in dar! k violet). Based in the CAD model a FLUKA geometry was created, only containing the geometry and the materi- als relevant for radiation calculations. Figure 3 shows the Z-X and the Z-Y views. Figure 1: 3D CAD view of the space below CBM service platform.
Figure 2: Very preliminary planning of rack positions in the service area.
Figure 2: Very preliminary planning of rack positions in the service area.
Figure 3: Upper panel: Room below the service platform in the X-Z plane at Y = - 5m, which is 0.7 m above the floor of the cave. Lower panel: Cross section in the Y-Z plane vertical to the beam.
Figure 3: Upper panel: Room below the service platform in the X-Z plane at Y = - 5m, which is 0.7 m above the floor of the cave. Lower panel: Cross section in the Y-Z plane vertical to the beam.
Figure 5: Ionizing dose after one month of beam integrated over Y from 0.5 to 1.0 m. Figure 4: Non-ionizing dose after one month of beam inte- grated over Y from 0.5 to 1.0 m.
Figure 5: Ionizing dose after one month of beam integrated over Y from 0.5 to 1.0 m. Figure 4: Non-ionizing dose after one month of beam inte- grated over Y from 0.5 to 1.0 m.
Figure 6: Hadron flux per cm2 and sec integrated over Y from 0.5 to 1.0 m.
Figure 6: Hadron flux per cm2 and sec integrated over Y from 0.5 to 1.0 m.
Figure 1: Dose rate distributions in the horizontal plane in the gap of the dipole magnet at different time steps after shutdown of the beam. The Au target was irradiated by an Au beam with an energy of 12A GeV and an intensity of 10° ions/s over one month. Figure | depicts the dose rate close to the target for dif- ferent times after shutdown of an Au beam which bom- barded the Au target with a kinetic energy of 12A GeV and an intensity of 10° ions/s over one month. Even one day after shut down, the dose rate in the magnet gap exceeds a value of 5 ywSv/h, in particular in front and inside the tar- get box. In figure 2 the dose rates are shown in a smaller scale, indicating that dose rate values of the order of 1000 Sv are reached in the close vicinity of the target still one day after beam shutdown. In order to reduce the radiation level when working at the target or close to the target box shortly after beam shutdown, a target shielding device is recommended. After removal of the target, the dose rate is drastically reduced as can be seen in figure 3.
Figure 1: Dose rate distributions in the horizontal plane in the gap of the dipole magnet at different time steps after shutdown of the beam. The Au target was irradiated by an Au beam with an energy of 12A GeV and an intensity of 10° ions/s over one month. Figure | depicts the dose rate close to the target for dif- ferent times after shutdown of an Au beam which bom- barded the Au target with a kinetic energy of 12A GeV and an intensity of 10° ions/s over one month. Even one day after shut down, the dose rate in the magnet gap exceeds a value of 5 ywSv/h, in particular in front and inside the tar- get box. In figure 2 the dose rates are shown in a smaller scale, indicating that dose rate values of the order of 1000 Sv are reached in the close vicinity of the target still one day after beam shutdown. In order to reduce the radiation level when working at the target or close to the target box shortly after beam shutdown, a target shielding device is recommended. After removal of the target, the dose rate is drastically reduced as can be seen in figure 3.
Figure 3: Dose rate distributions in the horizontal plane in the gap of the dipole magnet after removal of the target at different time steps after shutdown of the beam (see text). Figure 2: Dose rate distributions in the horizontal plane close to the target at different time steps after shutdown of the beam (see text).
Figure 3: Dose rate distributions in the horizontal plane in the gap of the dipole magnet after removal of the target at different time steps after shutdown of the beam (see text). Figure 2: Dose rate distributions in the horizontal plane close to the target at different time steps after shutdown of the beam (see text).
The GBT-based readout requires support for CROB boards capable to connect multiple FEBs, with different configurations of SMX2 chips. The structure of the DPB firmware was reorganized as shown in Fig. 1. The main Figure 1: Block diagram of the CROB-oriented DPB firmware (modified figure from [1]).
The GBT-based readout requires support for CROB boards capable to connect multiple FEBs, with different configurations of SMX2 chips. The structure of the DPB firmware was reorganized as shown in Fig. 1. The main Figure 1: Block diagram of the CROB-oriented DPB firmware (modified figure from [1]).
Figure 1: Block diagram of the GBTXEMU firmware. The block diagram of the GBTxEMU firmware is shown in Fig. 1. The reference clock is produced by the GTP re-
Figure 1: Block diagram of the GBTXEMU firmware. The block diagram of the GBTxEMU firmware is shown in Fig. 1. The reference clock is produced by the GTP re-
Figure 1: Overview of the long-haul test setup in its max- imum configuration using 2.4km of fiber. For shorter dis- tances fiber elements are removed from the link by chang- ing the patches. One goal for the new architecture is implementing the approximately 600m long connection between the FLES entry and processing clusters using standard InfiniBand network equipment if possible. This can significantly re- duce the cost for data transport compared to specialized long-haul equipment. Nevertheless in its default configura- tion InfiniBand EDR is only specified for reaches of 100 m. This limit is imposed by the available buffers needed to compensate the link delay and is defined by the bandwidth- delay product. Fortunately current InfiniBand switches allow to, at least partially, collapse buffers reserved for quality-of-service (QoS) features into a larger buffer at the cost of QoS. In theory this should increase the reach of the network sufficiently for CBM demands. To prove the fea- sibility of this concept a real world network end-to-end test was performed.
Figure 1: Overview of the long-haul test setup in its max- imum configuration using 2.4km of fiber. For shorter dis- tances fiber elements are removed from the link by chang- ing the patches. One goal for the new architecture is implementing the approximately 600m long connection between the FLES entry and processing clusters using standard InfiniBand network equipment if possible. This can significantly re- duce the cost for data transport compared to specialized long-haul equipment. Nevertheless in its default configura- tion InfiniBand EDR is only specified for reaches of 100 m. This limit is imposed by the available buffers needed to compensate the link delay and is defined by the bandwidth- delay product. Fortunately current InfiniBand switches allow to, at least partially, collapse buffers reserved for quality-of-service (QoS) features into a larger buffer at the cost of QoS. In theory this should increase the reach of the network sufficiently for CBM demands. To prove the fea- sibility of this concept a real world network end-to-end test was performed.
Figure 2: Possible minimal network architecture support- ing a total CBM data rate of 10.8 Tbit/s and up to 612 pro- cessing nodes. up to 2400 m in increments of 400 m and Mellanox PSM4 transceivers certified for up to 2km. Measurements for up to 2.4km were performed and show a very good agreement with theory. The maximum possible reach for one virtual lane, without loss of bandwidth has been calculated to be approximately 1020m. This is well sufficient for CBM’s needs. The available headroom suggests that also future In- finiBand generation may provide sufficient reach for CBM, even if buffer space is not increased proportionally to band- width.
Figure 2: Possible minimal network architecture support- ing a total CBM data rate of 10.8 Tbit/s and up to 612 pro- cessing nodes. up to 2400 m in increments of 400 m and Mellanox PSM4 transceivers certified for up to 2km. Measurements for up to 2.4km were performed and show a very good agreement with theory. The maximum possible reach for one virtual lane, without loss of bandwidth has been calculated to be approximately 1020m. This is well sufficient for CBM’s needs. The available headroom suggests that also future In- finiBand generation may provide sufficient reach for CBM, even if buffer space is not increased proportionally to band- width.
Figure 1: Data flow model of throttling system. The data flow in DAQ system is modeled as shown in Figure 1. The model is based on STS subsystem, involving the front-end delivering data and status information via the CRI to the TFC system, where busy status is evaluated and throttling decisions are taken and distributed.
Figure 1: Data flow model of throttling system. The data flow in DAQ system is modeled as shown in Figure 1. The model is based on STS subsystem, involving the front-end delivering data and status information via the CRI to the TFC system, where busy status is evaluated and throttling decisions are taken and distributed.
The model is verified though the regular event genera- tion. The model output can equal to input when the hit rate is up to maximum bandwidth. With the Poisson process, the model begins to lose data when the hit rate is more than about 98% of maximum bandwidth. The impact factors are throttling alert thresholds, channel disable time when throttling is on. The preliminary analysis is given as Fig- ure 3. The first element in each plot represents no throttling mechanism. With the throttling system, controlled loss in- creases obviously, while total loss remains a relative sta- ble level among the middle area. After this area, the total and controlled loss increase accordingly with longer chan- nel disable time, while uncontrolled loss nearly stays con- stant. Comparing two plots in Figure 3, lower alert thresh- olds lead to less uncontrolled loss than higher thresholds. These conclusions approve our expectation. Figure 3: Loss analysis with different channel disabling time.
The model is verified though the regular event genera- tion. The model output can equal to input when the hit rate is up to maximum bandwidth. With the Poisson process, the model begins to lose data when the hit rate is more than about 98% of maximum bandwidth. The impact factors are throttling alert thresholds, channel disable time when throttling is on. The preliminary analysis is given as Fig- ure 3. The first element in each plot represents no throttling mechanism. With the throttling system, controlled loss in- creases obviously, while total loss remains a relative sta- ble level among the middle area. After this area, the total and controlled loss increase accordingly with longer chan- nel disable time, while uncontrolled loss nearly stays con- stant. Comparing two plots in Figure 3, lower alert thresh- olds lead to less uncontrolled loss than higher thresholds. These conclusions approve our expectation. Figure 3: Loss analysis with different channel disabling time.
Figure 1: FTLMC second protype @mCBM.
Figure 1: FTLMC second protype @mCBM.
Figure 1: The ECS prototype architecture features a three- level hierarchy with support for independent partitions
Figure 1: The ECS prototype architecture features a three- level hierarchy with support for independent partitions
Figure 2: Example of a detector subsystem state-machine
Figure 2: Example of a detector subsystem state-machine
In order to increase the throughput, FLESnet needs to complete timeslices more rapidly to free the buffer space. Figure 2 depicts the duration between the arrival of the first and the last contribution of each timeslice. It shows that DFS requires a stable clock to complete timeslices in a timely manner. Figure 3 depicts the variance of complet- ing a timeslice after excluding the warm-up phase, which is the first few intervals. When the clocks are synchronized, DFS reduces the duration to complete timeslices by up to 90 %. Figure 2: Duration between first and last contribution of timeslice
In order to increase the throughput, FLESnet needs to complete timeslices more rapidly to free the buffer space. Figure 2 depicts the duration between the arrival of the first and the last contribution of each timeslice. It shows that DFS requires a stable clock to complete timeslices in a timely manner. Figure 3 depicts the variance of complet- ing a timeslice after excluding the warm-up phase, which is the first few intervals. When the clocks are synchronized, DFS reduces the duration to complete timeslices by up to 90 %. Figure 2: Duration between first and last contribution of timeslice
Figure 3: 10th-90th percentile of duration to receive times- lices
Figure 3: 10th-90th percentile of duration to receive times- lices
Table 1: Event finding efficiency.
Table 1: Event finding efficiency.
Simple event builder developed for the CBM should op- erate at digi level. The event building procedure consists of two steps: finding of events in the data stream and as- signing digis to the found event (i.e. event composition). In the 2018 we have adopted algorithms, developed for full SIS100 CBM setup, for the mCBM experiment. Figure 1: mCBM model used for event building. Beam comes from the right.
Simple event builder developed for the CBM should op- erate at digi level. The event building procedure consists of two steps: finding of events in the data stream and as- signing digis to the found event (i.e. event composition). In the 2018 we have adopted algorithms, developed for full SIS100 CBM setup, for the mCBM experiment. Figure 1: mCBM model used for event building. Beam comes from the right.
gle event for STS, MUCH, TRD and TOF are shown in Fig. 2. The distribution for TRD looks odd and should be understood. The width of the distribution cores for STS, MUCH and TOF are similar. Long tails are mostly due to the delta electrons and can be discarded without negative impact to the physical performance. Same extended time period can be used for event composition in all three detec- tors. Time-of-flight and electronics delays shifts should be corrected. 99.2% of MUCH and 99.3% of TOF digis are correctly attached to the found events after all corrections. Figure 2: Digi time for a single event. STS digis are shown in top left plot, MUCH — in top right, TDR — in bottom left and TOF in bottom right.
gle event for STS, MUCH, TRD and TOF are shown in Fig. 2. The distribution for TRD looks odd and should be understood. The width of the distribution cores for STS, MUCH and TOF are similar. Long tails are mostly due to the delta electrons and can be discarded without negative impact to the physical performance. Same extended time period can be used for event composition in all three detec- tors. Time-of-flight and electronics delays shifts should be corrected. 99.2% of MUCH and 99.3% of TOF digis are correctly attached to the found events after all corrections. Figure 2: Digi time for a single event. STS digis are shown in top left plot, MUCH — in top right, TDR — in bottom left and TOF in bottom right.
In this report the status of global track reconstruction with the Cellular Automaton (CA) track finder package in the CBM experiment is presented. The proposed proce- dure in the muon setup of the CBM detector, i.e. with MUCH (Muon Chamber) system, is based on the extended CA method, which is the standard reconstruction routine for the main tracking detector of CBM — Silicon Tracking System (STS). Figure 1: Schematic view of the combined system STS and MUCH SIS100-B configuration. MUCH consists of 12 de- tector layers and 4 absorbers including the first absorber of 60 cm carbon.
In this report the status of global track reconstruction with the Cellular Automaton (CA) track finder package in the CBM experiment is presented. The proposed proce- dure in the muon setup of the CBM detector, i.e. with MUCH (Muon Chamber) system, is based on the extended CA method, which is the standard reconstruction routine for the main tracking detector of CBM — Silicon Tracking System (STS). Figure 1: Schematic view of the combined system STS and MUCH SIS100-B configuration. MUCH consists of 12 de- tector layers and 4 absorbers including the first absorber of 60 cm carbon.
Table 1: Track reconstruction performance for central UrQMD AuAu collisons at 1|OAGeV beam energy with two embedded muons per event: efficiencies (%) for different sets of tracks, clone and ghost levels (%). Goethe-Universitat, Frankfurt am Main, Germany
Table 1: Track reconstruction performance for central UrQMD AuAu collisons at 1|OAGeV beam energy with two embedded muons per event: efficiencies (%) for different sets of tracks, clone and ghost levels (%). Goethe-Universitat, Frankfurt am Main, Germany
Figure 2: Reconstructed m? of found tracks as a function of their momentum. Blue symbols represent muons from w-decays.
Figure 2: Reconstructed m? of found tracks as a function of their momentum. Blue symbols represent muons from w-decays.
Figure 1: MC track ID of the reconstructed tracks from the vector finder and /ittrack packages: 0 and 1 correspond to the MUCH reconstructed tracks of the muons from w- decays correctly matched to the muon tracks in the STS, 5 represents all the other tracks. m? of the found tracks is plotted versus their momentum with m? = (t?c? /I?-1)p?, t is the time measurement from the TOF, p is the momentum from the STS, / is the distance from the target to the TOF hit and c is the speed of light. One can see, that the quality of track matching of all de- tectors allows one to apply a selection window in m? for signal muons.
Figure 1: MC track ID of the reconstructed tracks from the vector finder and /ittrack packages: 0 and 1 correspond to the MUCH reconstructed tracks of the muons from w- decays correctly matched to the muon tracks in the STS, 5 represents all the other tracks. m? of the found tracks is plotted versus their momentum with m? = (t?c? /I?-1)p?, t is the time measurement from the TOF, p is the momentum from the STS, / is the distance from the target to the TOF hit and c is the speed of light. One can see, that the quality of track matching of all de- tectors allows one to apply a selection window in m? for signal muons.
Figure 1: Green tracks are reconstructable (can be used for the alignment), where red tracks aren’t. To be able to align the detector components (alignables) successfully using a track based alignment algorithm, the alignables should have a decent number of reconstructed tracks (at least 4 STS hits to a track) passing through them. For straight tracks, originating from the target, some lad- ders located at large angles have not enough overlap with other ladders for a meaningful track based alignment strat- egy (see Fig. 1).
Figure 1: Green tracks are reconstructable (can be used for the alignment), where red tracks aren’t. To be able to align the detector components (alignables) successfully using a track based alignment algorithm, the alignables should have a decent number of reconstructed tracks (at least 4 STS hits to a track) passing through them. For straight tracks, originating from the target, some lad- ders located at large angles have not enough overlap with other ladders for a meaningful track based alignment strat- egy (see Fig. 1).
Figure 2: Track slope distribution in X direction (J;,) is in green and in Y direction (Z;,) is in red with respect to the beam axis (no unit). purpose a simple cosmic ray generator is developed. One million cosmic straight muon tracks with a fixed momen- tum of 5Gev/c are simulated without the magnetic field.
Figure 2: Track slope distribution in X direction (J;,) is in green and in Y direction (Z;,) is in red with respect to the beam axis (no unit). purpose a simple cosmic ray generator is developed. One million cosmic straight muon tracks with a fixed momen- tum of 5Gev/c are simulated without the magnetic field.
Figure 3: Residual dist. in X direction.
Figure 3: Residual dist. in X direction.
Figure 4: Residual dist. in Y direction.
Figure 4: Residual dist. in Y direction.
Figure 5: X directional misalignment correction.
Figure 5: X directional misalignment correction.
Figure 6: Y directional misalignment correction.
Figure 6: Y directional misalignment correction.
Figure 1: Efficiency corrected spectra of = — Am de- cay as a function of m; obtained by the conventional (left) and missing mass (right) methods from 5M central AuAu UrQMD events at 10 AGeV. For extraction of signal spec- tra two methods were used for each case: the side bands (blue) and polynomial background fit (green) methods, which are shown in comparison with the simulated MC sig- nal (red). Slope is extracted from the exponential fit.
Figure 1: Efficiency corrected spectra of = — Am de- cay as a function of m; obtained by the conventional (left) and missing mass (right) methods from 5M central AuAu UrQMD events at 10 AGeV. For extraction of signal spec- tra two methods were used for each case: the side bands (blue) and polynomial background fit (green) methods, which are shown in comparison with the simulated MC sig- nal (red). Slope is extracted from the exponential fit.
Figure 1: ©~ mass (left) and y? probability (right) distri- bution spectra which now has correct flat shape (red) com- paring to the previous version (blue).
Figure 1: ©~ mass (left) and y? probability (right) distri- bution spectra which now has correct flat shape (red) com- paring to the previous version (blue).
To recover the efficiency of particle identification, tracks without PID are considered as pions. Since pions are abun- dant particles, this step increases the efficiency of recon- structing particles with pion as a daughter, but does not af- fect the significance of the signal. As for the decay with protons and kaons in the decay channel, this compensation leads to a large increase in the background, so it is only applied to pions. Figure 2: S~ > m7 +n (left) and ©+ > 7° +p (right) ef- ficiency comparison for old (blue) and new (red) versions. The increase of significance is 40% and 10% correspond- ingly using 5M centr Urqmd Au+Au events at 10 AGeV with TOF PID.
To recover the efficiency of particle identification, tracks without PID are considered as pions. Since pions are abun- dant particles, this step increases the efficiency of recon- structing particles with pion as a daughter, but does not af- fect the significance of the signal. As for the decay with protons and kaons in the decay channel, this compensation leads to a large increase in the background, so it is only applied to pions. Figure 2: S~ > m7 +n (left) and ©+ > 7° +p (right) ef- ficiency comparison for old (blue) and new (red) versions. The increase of significance is 40% and 10% correspond- ingly using 5M centr Urqmd Au+Au events at 10 AGeV with TOF PID.
Figure 2: Schematic view of using multiple inputs for dig- itization read-out electronics. Such sources are usually addressed by separate transport simulations, which have to be simul- taneously fed into the digitization procedure with differ- ent rates (’mixing“). In addition, the possibility to em- bed several transport data sources at digitization level is highly desirable, e.g., embedding a rare signal into full background events. This embedding was previously only possible at transport level by transporting signal and back- ground together. As the transport simulation is the most compute-intensive stage, it is economically preferable to transport signal and background separately, such that the high-multiplicity background events can be re-used for a variety of signals. In contrast to mixing data inputs, where the resulting data stream is composed of events randomly chosen from one input or the other according to the respec- tive event rates, embedding means a one-to-one correspon- dence of two events from different sources, i.e., both event times have to be equal. These demands were met by the introduction of the class CbmDigitizationSource covering the above described functionality. The working principle is demonstrated in Fig. 2: several MC (transport) inputs can be mutually em- bedded into one ’input set“. All inputs within an input set will be overlaid by assigning the same event time. Several input sets can be mixed with different event rates. The dig- itization procedure will for each input set sample the event time according to the specified rate and process the events in a time-ordered manner.
Figure 2: Schematic view of using multiple inputs for dig- itization read-out electronics. Such sources are usually addressed by separate transport simulations, which have to be simul- taneously fed into the digitization procedure with differ- ent rates (’mixing“). In addition, the possibility to em- bed several transport data sources at digitization level is highly desirable, e.g., embedding a rare signal into full background events. This embedding was previously only possible at transport level by transporting signal and back- ground together. As the transport simulation is the most compute-intensive stage, it is economically preferable to transport signal and background separately, such that the high-multiplicity background events can be re-used for a variety of signals. In contrast to mixing data inputs, where the resulting data stream is composed of events randomly chosen from one input or the other according to the respec- tive event rates, embedding means a one-to-one correspon- dence of two events from different sources, i.e., both event times have to be equal. These demands were met by the introduction of the class CbmDigitizationSource covering the above described functionality. The working principle is demonstrated in Fig. 2: several MC (transport) inputs can be mutually em- bedded into one ’input set“. All inputs within an input set will be overlaid by assigning the same event time. Several input sets can be mixed with different event rates. The dig- itization procedure will for each input set sample the event time according to the specified rate and process the events in a time-ordered manner.
The standard Monte Carlo STS geometry (v16g, Fig. 1) for the CBM ROOT is very detailed and includes all active and passive elements in the detector acceptance relevant for the particle track reconstruction, e.g. silicon sensors, read- out cables and carbon-fiber ladders with the proper material budget. Figure 1: A render of a current standard STS geometry version v1l6g.
The standard Monte Carlo STS geometry (v16g, Fig. 1) for the CBM ROOT is very detailed and includes all active and passive elements in the detector acceptance relevant for the particle track reconstruction, e.g. silicon sensors, read- out cables and carbon-fiber ladders with the proper material budget. Figure 1: A render of a current standard STS geometry version v1l6g.
Figure 3: A render of new STS geometry version v19a with passive materials. A hybrid approach was established for producing the fi- nal ROOT geometry. The ladders with sensors are pro- duced in the well established scripted manner with many steering variables. The passive materials are imported from a GDML file and added to geometry dynamically (Fig 3).
Figure 3: A render of new STS geometry version v19a with passive materials. A hybrid approach was established for producing the fi- nal ROOT geometry. The ladders with sensors are pro- duced in the well established scripted manner with many steering variables. The passive materials are imported from a GDML file and added to geometry dynamically (Fig 3).
Figure 2: A render of the Unit 4 imported into ROOT. FAIR, Darmstadt, Germany
Figure 2: A render of the Unit 4 imported into ROOT. FAIR, Darmstadt, Germany
Carbon and divided into two parts. The first part is 16 cm thick and tra pezoidal in shape followed by a 44 cm thick parallelepiped. Rest of the absorbers are made up of Iron and paral elepiped in shape. In present simulation two variants of MuCh geometry (tag version- v17b_lmvm and vi9a_lmvm) have been studied. The v17b_lmvm ge- ometry contains .5 cm rectangular plates made of Carbon plastic as a support structure but no arrangement for cool- ing. In the modified version (v19a_1mvm) carbon plates are replaced by 1 cm thick Aluminum plates for support and cooling purpose and to make the geometry more realistic we also implemented 35 micron Copper and 3 mm G10 on both sides of active volume of the detector for all stations. The internal structure of each module as per the latest ge- ometry concern are shown in Fig. 1. Figure 1: Internal structure of each module in realistic MuCh geometry (v19a_1lmvm).
Carbon and divided into two parts. The first part is 16 cm thick and tra pezoidal in shape followed by a 44 cm thick parallelepiped. Rest of the absorbers are made up of Iron and paral elepiped in shape. In present simulation two variants of MuCh geometry (tag version- v17b_lmvm and vi9a_lmvm) have been studied. The v17b_lmvm ge- ometry contains .5 cm rectangular plates made of Carbon plastic as a support structure but no arrangement for cool- ing. In the modified version (v19a_1mvm) carbon plates are replaced by 1 cm thick Aluminum plates for support and cooling purpose and to make the geometry more realistic we also implemented 35 micron Copper and 3 mm G10 on both sides of active volume of the detector for all stations. The internal structure of each module as per the latest ge- ometry concern are shown in Fig. 1. Figure 1: Internal structure of each module in realistic MuCh geometry (v19a_1lmvm).
Figure 2: Point density of primary(top) and sec- ondary(bottom) particles for each MuCh station.
Figure 2: Point density of primary(top) and sec- ondary(bottom) particles for each MuCh station.
Figure 3: Invariant Mass Spectra for the Combinatorial Background.
Figure 3: Invariant Mass Spectra for the Combinatorial Background.
The readout planes of the modules are segmented in pads for obtaining final detectable response. The procedure of distributing the MUCH points to pads, known as digitiza- tion involves the detailed procedure of implementing the response of the gas detector to the energy deposition inside the chamber.
The readout planes of the modules are segmented in pads for obtaining final detectable response. The procedure of distributing the MUCH points to pads, known as digitiza- tion involves the detailed procedure of implementing the response of the gas detector to the energy deposition inside the chamber.
Figure 1: Arrangement of modules in staggerred manner on both sides of a carbon support structure. RPC module has been shown separately. In simulation, for the 1° & 2”¢ MUCH-stations, GEM detector with 3mm Argon based gas mixture has been used. For 3™¢ & 4" station, at the moment, we have consid- ered sector shaped modules for RPC, similar to that used in MUCH-GEM. Each module consists of 2mm RPC gas sandwiched between 2mm RPC glass electrodes on both sides. Modules are arranged in staggerred manners on both sides of a carbon support structure as shown in Fig. 1. RPC Gas composition : TetraFluorethan (C2H2F,) 85%, Sulfurhexafluoride (SFg) 10% & 5% Isobutane (C4Hi0), presently used in simulation.
Figure 1: Arrangement of modules in staggerred manner on both sides of a carbon support structure. RPC module has been shown separately. In simulation, for the 1° & 2”¢ MUCH-stations, GEM detector with 3mm Argon based gas mixture has been used. For 3™¢ & 4" station, at the moment, we have consid- ered sector shaped modules for RPC, similar to that used in MUCH-GEM. Each module consists of 2mm RPC gas sandwiched between 2mm RPC glass electrodes on both sides. Modules are arranged in staggerred manners on both sides of a carbon support structure as shown in Fig. 1. RPC Gas composition : TetraFluorethan (C2H2F,) 85%, Sulfurhexafluoride (SFg) 10% & 5% Isobutane (C4Hi0), presently used in simulation.
MUCH covers an angular region from 5.7° to 25°. Each station has three detector layers. Detector layers are of cir- cular shape & consists of several trapezoidal modules. In 1°! & 2"4 station azimuthal 1° segmentation on pads has been used. 1** station has 16 modules, 2”@ station has 20 modules. For 3"¢ & 4*” station 5° & 6° segmentation have been implemented as the track density is much less there and accordingly 18 & 20 modules have been used respec- tively as shown in Fig. 2. Figure 2: 5° (left fig) & 6° (right fig) azimuthal segmenta- tion in station 3 & station 4 respectively.
MUCH covers an angular region from 5.7° to 25°. Each station has three detector layers. Detector layers are of cir- cular shape & consists of several trapezoidal modules. In 1°! & 2"4 station azimuthal 1° segmentation on pads has been used. 1** station has 16 modules, 2”@ station has 20 modules. For 3"¢ & 4*” station 5° & 6° segmentation have been implemented as the track density is much less there and accordingly 18 & 20 modules have been used respec- tively as shown in Fig. 2. Figure 2: 5° (left fig) & 6° (right fig) azimuthal segmenta- tion in station 3 & station 4 respectively.
Figure 5: Radial hit density distributions at different sta- tions. Figure 4: Radial point density distributions at different sta- tions.
Figure 5: Radial hit density distributions at different sta- tions. Figure 4: Radial point density distributions at different sta- tions.
The radial distribution of point density & hit density for 3"¢ & 4'” station have been shown in Fig. 4 and Fig. 5 respectively. These are somewhat lower compared to pre- vious GEM geometry and digitization [2].
The radial distribution of point density & hit density for 3"¢ & 4'” station have been shown in Fig. 4 and Fig. 5 respectively. These are somewhat lower compared to pre- vious GEM geometry and digitization [2].
Figure 3: Number of primary electrons distribution within RPC drift volume. With these realististic geometry and digitization param- eters we have simulated the performance of MUCH. This simulation have been done with SIS1O0B (4 stations + 4 absorbers) geometry at 8 AGeV central Au+Au collisions using UrQMD and PLUTO embedded events. The primary electron distribution from each MC point, within RPC de- tector drift volume has been shown in Fig. 3.
Figure 3: Number of primary electrons distribution within RPC drift volume. With these realististic geometry and digitization param- eters we have simulated the performance of MUCH. This simulation have been done with SIS1O0B (4 stations + 4 absorbers) geometry at 8 AGeV central Au+Au collisions using UrQMD and PLUTO embedded events. The primary electron distribution from each MC point, within RPC de- tector drift volume has been shown in Fig. 3.
Figure 7: Radial distribution of occupancy for station 4. Figure 6: Radial distribution of occupancy for station 3.
Figure 7: Radial distribution of occupancy for station 4. Figure 6: Radial distribution of occupancy for station 3.
Figure 8: Points/cluster and digis/cluster distribution at sta- tion 3 and 4.
Figure 8: Points/cluster and digis/cluster distribution at sta- tion 3 and 4.
For this purpose we estimate MuCh data rate by us- ing event coherent background source from GEANT3 and GEANT4. First to see the difference in effect of GEANT3 and GEANT4 transport models, we analyse 10° minimum bias Au-Au events from UrQMD at beam energy 12A GeV. The produced particles are passed through the SIS- 100 muon setup which includes 5 hadron absorbers and 4 tracking stations. Each station consists of three GEM lay- ers. We compare point density/event for each station of MuCh for GEANT3 and GEANT4. It is observed that the contribution of secondary particles has been increased us- ing GEANT4 compared to GEANT3. There is hardly any change in the contribution of primary particles as expected. It can be easily seen from Table 1. Table 1: Difference in production of primary and sec- ondary particles using GEANT3 and GEANT4.
For this purpose we estimate MuCh data rate by us- ing event coherent background source from GEANT3 and GEANT4. First to see the difference in effect of GEANT3 and GEANT4 transport models, we analyse 10° minimum bias Au-Au events from UrQMD at beam energy 12A GeV. The produced particles are passed through the SIS- 100 muon setup which includes 5 hadron absorbers and 4 tracking stations. Each station consists of three GEM lay- ers. We compare point density/event for each station of MuCh for GEANT3 and GEANT4. It is observed that the contribution of secondary particles has been increased us- ing GEANT4 compared to GEANT3. There is hardly any change in the contribution of primary particles as expected. It can be easily seen from Table 1. Table 1: Difference in production of primary and sec- ondary particles using GEANT3 and GEANT4.
In CBMROOT simulations, particles generated via event generator are made to pass through the absorbers and de- tector layers of MuCh using GEANT3 transport engine. But GEANT3 seems to underestimate particle rate, partic- ularly those produced due to nuclear interaction. So we used GEANT4 instead of GEANT3 because GEANT4 in- cludes some additional process that represents the interac- tion more reasonably compared to GEANT3. These phe- nomena inside the most recent version of GEANT4 provide good agreement of energy response, resolution of pions and protons.
In CBMROOT simulations, particles generated via event generator are made to pass through the absorbers and de- tector layers of MuCh using GEANT3 transport engine. But GEANT3 seems to underestimate particle rate, partic- ularly those produced due to nuclear interaction. So we used GEANT4 instead of GEANT3 because GEANT4 in- cludes some additional process that represents the interac- tion more reasonably compared to GEANT3. These phe- nomena inside the most recent version of GEANT4 provide good agreement of energy response, resolution of pions and protons.
Table 2: Maximum foreseen pad hit rate due to event co- herent background source, for 12A GeV Au-Au collision for each station using GEANT3 and GEANT4.
Table 2: Maximum foreseen pad hit rate due to event co- herent background source, for 12A GeV Au-Au collision for each station using GEANT3 and GEANT4.
Figure 2: Occupancy for different MuCh stations using GEANT3 and GEANT4 transport models.
Figure 2: Occupancy for different MuCh stations using GEANT3 and GEANT4 transport models.
Figure 3: Comparison of occupancy for Ist station at dif- ferent beam energies.
Figure 3: Comparison of occupancy for Ist station at dif- ferent beam energies.
Figure 2: Web-interface: Edit Mode icon in Available Setups Edit Mode (see Figure 2). It is needed to confirm the editing of setup in the popup win- dow. After that one can modify the setup structure. The CBM Geometry DB has been developed for operat- ing with information about the geometry of the CBM setup organized in FairRoot format. This means that it can be used by other physical experiments based on the ROOT en- vironment.
Figure 2: Web-interface: Edit Mode icon in Available Setups Edit Mode (see Figure 2). It is needed to confirm the editing of setup in the popup win- dow. After that one can modify the setup structure. The CBM Geometry DB has been developed for operat- ing with information about the geometry of the CBM setup organized in FairRoot format. This means that it can be used by other physical experiments based on the ROOT en- vironment.
A comparison of the reconstruction performance for each of the three identification approaches, in combina- tion with different sets of cuts, is presented in Figure 2. The correlation between points has roughly exponential de- pendence (red curvature is a fit of data with an exponential function). Figure 2: Correlation between reconstructed number of 77 and signal to background ratio.
A comparison of the reconstruction performance for each of the three identification approaches, in combina- tion with different sets of cuts, is presented in Figure 2. The correlation between points has roughly exponential de- pendence (red curvature is a fit of data with an exponential function). Figure 2: Correlation between reconstructed number of 77 and signal to background ratio.
lysis, three different approaches are studied: full ID (4 out 4 are identified), partial ID (at least one lepton out of each pair is identified), without ID (0 or more out of 4 are iden- tified). An example of reconstructed results can be seen in Figure 1. Figure 1: Reconstructed invariant mass spectrum of 7) after background subtraction using double conversion method.
lysis, three different approaches are studied: full ID (4 out 4 are identified), partial ID (at least one lepton out of each pair is identified), without ID (0 or more out of 4 are iden- tified). An example of reconstructed results can be seen in Figure 1. Figure 1: Reconstructed invariant mass spectrum of 7) after background subtraction using double conversion method.
All simulations have been performed in CbmRoot us- ing the dielectron analysis package. For both setups a tar- get thickness of 25 jzm has been chosen. The magnetic field has been scaled to 50% for Au+Au collisions and to 60% for Ag+Ag collisions. All CBM detectors were imple- mented using the standard geometries as of Summer 2018 except the MVD detector. Simulations were performed us- ing UrQMD as event generator. In each event a signal di- electron pair generated by PLUTO was embedded. As sig- nal di-electrons either w, ¢ and in-medium p [4] decays into e+ e~, w-Dalitz decays, or QGP radiation was imple- mented. Signals were scaled later according to their esti- mated multiplicity and branching fraction [5]. It is impor- tant to notice, that scaling factors were changed from the previously used estimates from HSD to values using a ther- mal model [6]. This reduced the w multiplicity by a factor 8 but enlarged the @ multiplicity by a factor 3. Thus com- parisons of S/B ratios from these simulations to previous simulations have to account for this factor. Figure 1: Invariant mass spectra for 5M central Au+Au col- lisions (left) at 8 AGeV and 10M minbias Ag+Ag collisions (right) at 4.5 AGeV beam energy excluding any pt-cut.
All simulations have been performed in CbmRoot us- ing the dielectron analysis package. For both setups a tar- get thickness of 25 jzm has been chosen. The magnetic field has been scaled to 50% for Au+Au collisions and to 60% for Ag+Ag collisions. All CBM detectors were imple- mented using the standard geometries as of Summer 2018 except the MVD detector. Simulations were performed us- ing UrQMD as event generator. In each event a signal di- electron pair generated by PLUTO was embedded. As sig- nal di-electrons either w, ¢ and in-medium p [4] decays into e+ e~, w-Dalitz decays, or QGP radiation was imple- mented. Signals were scaled later according to their esti- mated multiplicity and branching fraction [5]. It is impor- tant to notice, that scaling factors were changed from the previously used estimates from HSD to values using a ther- mal model [6]. This reduced the w multiplicity by a factor 8 but enlarged the @ multiplicity by a factor 3. Thus com- parisons of S/B ratios from these simulations to previous simulations have to account for this factor. Figure 1: Invariant mass spectra for 5M central Au+Au col- lisions (left) at 8 AGeV and 10M minbias Ag+Ag collisions (right) at 4.5 AGeV beam energy excluding any pt-cut.
Figure 2: Efficiencies of the w-meson reconstruction in the y — pt plane compared to MC input for central Au+Au col- lisions (left) and minimum bias Ag+Ag collisions (right) excluding any pt-cut. Numbers in the upper left corner are total integrated efficiencies. Midrapidity is at 1.4 for 8 AGeV and 1.1 for 4.5 AGeV beam energy. 'Justus-Liebig University Giessen, Giessen, Germany; 2LIT JINR, Dubna, Russia; *GSI, Darmstadt, Germany
Figure 2: Efficiencies of the w-meson reconstruction in the y — pt plane compared to MC input for central Au+Au col- lisions (left) and minimum bias Ag+Ag collisions (right) excluding any pt-cut. Numbers in the upper left corner are total integrated efficiencies. Midrapidity is at 1.4 for 8 AGeV and 1.1 for 4.5 AGeV beam energy. 'Justus-Liebig University Giessen, Giessen, Germany; 2LIT JINR, Dubna, Russia; *GSI, Darmstadt, Germany
Figure 1: (top) vi(pr) for r+ for centrality 10-35% for Poeam = 10A GeV/c. (middle) v1 (y) for A and K® for centrality 20-30% for Prcam = 12A GeV/c. (bottom) v1 (y) for A for Ppeam = 3.3A GeV/c. ‘National Research Center Kurchatov Institute”, Moscow, Russia; 2National Research Nuclear University MEPhI (Moscow Engineering Physics Institute), Moscow, Russia; 3GSI Helmholtzzentrum fiir Schwerionenforschung, Darmstadt, Germany; 4Goethe University Frankfurt, Frankfurt am Main, Germany
Figure 1: (top) vi(pr) for r+ for centrality 10-35% for Poeam = 10A GeV/c. (middle) v1 (y) for A and K® for centrality 20-30% for Prcam = 12A GeV/c. (bottom) v1 (y) for A for Ppeam = 3.3A GeV/c. ‘National Research Center Kurchatov Institute”, Moscow, Russia; 2National Research Nuclear University MEPhI (Moscow Engineering Physics Institute), Moscow, Russia; 3GSI Helmholtzzentrum fiir Schwerionenforschung, Darmstadt, Germany; 4Goethe University Frankfurt, Frankfurt am Main, Germany
The analyzes in this article were performed using the UrQMD 3.3 generator, integrated in the YaPT system, with t = 200 fm/c, for which the state equation is defined by the cascade mode, and the AMPT 2.26t7 generator, which uses the string fragmentation Lund model and the pop- corn mechanism (baryon stopping). AMPT shows an equal ammount of strangeness for the 10-11 and 12-13 A GeV in- tervals. This equality may suggest a pre-equilibrium in the highlited intervals, especially for the 0 < y < 0.8 interval, where we can observe a liniar increase of the ratios, under and above the 10-13 A GeV interval, for both centralities (for = xv: Fig. 1). Figure 1: amer Re ratio distribution for two centralities, considering different collision energies and two rapidity intervals: left)O << y < 0.8;right)O.5 << y < 1.4
The analyzes in this article were performed using the UrQMD 3.3 generator, integrated in the YaPT system, with t = 200 fm/c, for which the state equation is defined by the cascade mode, and the AMPT 2.26t7 generator, which uses the string fragmentation Lund model and the pop- corn mechanism (baryon stopping). AMPT shows an equal ammount of strangeness for the 10-11 and 12-13 A GeV in- tervals. This equality may suggest a pre-equilibrium in the highlited intervals, especially for the 0 < y < 0.8 interval, where we can observe a liniar increase of the ratios, under and above the 10-13 A GeV interval, for both centralities (for = xv: Fig. 1). Figure 1: amer Re ratio distribution for two centralities, considering different collision energies and two rapidity intervals: left)O << y < 0.8;right)O.5 << y < 1.4
Figure 2: uremp and two rapidity intervals: left)O << y < 0.8:right)0.5 < y < 1.4 ratio distribution for two centralities, considering different collision energies Figure 3: UrQMD B ratio distribution for two centralities, considering different collision energies and :left)O < y < O.8;right)0.5 < y < 1.4 wo rapidiy interval
Figure 2: uremp and two rapidity intervals: left)O << y < 0.8:right)0.5 < y < 1.4 ratio distribution for two centralities, considering different collision energies Figure 3: UrQMD B ratio distribution for two centralities, considering different collision energies and :left)O < y < O.8;right)0.5 < y < 1.4 wo rapidiy interval
motivated to investigate 1 A GeV intervals for the excita- tion function and even 0.5 A GeV, such that we can see the formation or ”kind” of the plateau. We can also ob- serve a peak at 18 A GeV (Fig. 3, left). These plateaus and the peak suggest a possible mixed phase. The plateau forms that we obtained seems to be in agreement with the results from [1]. Here, the results suggest the formation of the mixed phase based on the calculations of pion/baryon, and entropy/baryon ratios in the energy range 6.9 - 11.6 A GeV, and these are compared with the AGS (Au-Au) and SPS (Pb-Pb) data for regions with anomalous thermody- namic properties.
motivated to investigate 1 A GeV intervals for the excita- tion function and even 0.5 A GeV, such that we can see the formation or ”kind” of the plateau. We can also ob- serve a peak at 18 A GeV (Fig. 3, left). These plateaus and the peak suggest a possible mixed phase. The plateau forms that we obtained seems to be in agreement with the results from [1]. Here, the results suggest the formation of the mixed phase based on the calculations of pion/baryon, and entropy/baryon ratios in the energy range 6.9 - 11.6 A GeV, and these are compared with the AGS (Au-Au) and SPS (Pb-Pb) data for regions with anomalous thermody- namic properties.
Figure 1: CBM geometry for muon measurements: STS+MuCh+TRD+ToF. Muon system (MuCh) compris- ing 2 GEM stations with 1° radial segmentation and 3 ab- sorbers: 60 cm carbon and 20+20 cm iron. This report is based on the CBM Internal Physics Note 18001. In this note we propose a start version of the Muon Chamber (MuCh) system which will be well suited for the identification of muons in heavy-ion collisions at kinetic beam energies up to about 4A GeV. The proposal is based on a MuCh comprising 2 GEM stations and 3 absorbers only. The radial segmentation of the two GEM detectors is 1°, the absorbers consist of a Carbon block of 60 cm thick- ness in front of the first GEM station, a 20 cm thick iron layer between the first and the second GEM station, and a 20 cm iron layer behind the second GEM station. In addi- tion, the experimental setup includes the Silicon Tracking Stations (STS) for track reconstruction, the Time-Of-Flight wall (TOF) for particle identification, and the Transition Radiation Detector (TRD) as intermediate tracker between the second GEM station and the TOF detector, as illustrated in figure 1.
Figure 1: CBM geometry for muon measurements: STS+MuCh+TRD+ToF. Muon system (MuCh) compris- ing 2 GEM stations with 1° radial segmentation and 3 ab- sorbers: 60 cm carbon and 20+20 cm iron. This report is based on the CBM Internal Physics Note 18001. In this note we propose a start version of the Muon Chamber (MuCh) system which will be well suited for the identification of muons in heavy-ion collisions at kinetic beam energies up to about 4A GeV. The proposal is based on a MuCh comprising 2 GEM stations and 3 absorbers only. The radial segmentation of the two GEM detectors is 1°, the absorbers consist of a Carbon block of 60 cm thick- ness in front of the first GEM station, a 20 cm thick iron layer between the first and the second GEM station, and a 20 cm iron layer behind the second GEM station. In addi- tion, the experimental setup includes the Silicon Tracking Stations (STS) for track reconstruction, the Time-Of-Flight wall (TOF) for particle identification, and the Transition Radiation Detector (TRD) as intermediate tracker between the second GEM station and the TOF detector, as illustrated in figure 1.
Figure 2: Reconstructed muon tracks from omega meson decays (left panel) and reconstructed background tracks (right panel) as function of mass squared and momentum based on TOF information for central Au+Au collisions with beam momentum of 4 A GeV/c. The black lines il- lustrate the selection criteria for muons. performed for Au beams with momenta of 2 and 4 A GeV/c, a 1% interaction Au target, and a magnetic field o 50% of the maximum va are generated using the into central Au+Au eve code. In order to identi have been applied: primary vertex cut (y < 2), number o hits in STS (> 7), number of hits in MuCh (> 5), num- ue. The low-mass vector mesons PLUTO code, and are embedded nts generated with the UrQMD fy the muons, the following cuts ber of hits in TRD (> MuCh (x < 2). Finally, ), and track quality in STS and a 20 cut has been applied in the plane mass-squared versus momentum provided by TOF, as illustrated in figure 2.
Figure 2: Reconstructed muon tracks from omega meson decays (left panel) and reconstructed background tracks (right panel) as function of mass squared and momentum based on TOF information for central Au+Au collisions with beam momentum of 4 A GeV/c. The black lines il- lustrate the selection criteria for muons. performed for Au beams with momenta of 2 and 4 A GeV/c, a 1% interaction Au target, and a magnetic field o 50% of the maximum va are generated using the into central Au+Au eve code. In order to identi have been applied: primary vertex cut (y < 2), number o hits in STS (> 7), number of hits in MuCh (> 5), num- ue. The low-mass vector mesons PLUTO code, and are embedded nts generated with the UrQMD fy the muons, the following cuts ber of hits in TRD (> MuCh (x < 2). Finally, ), and track quality in STS and a 20 cut has been applied in the plane mass-squared versus momentum provided by TOF, as illustrated in figure 2.
Figure 4: Reconstructed dimuon invariant mass distribu- tion simulated for central Au+Au collisions with a beam momentum of 4 GeV/c (black histogram). The dimuon decays of low-mass vector mesons including their Dalitz decays, which have been identified using Monte Carlo par- ticle information, are shown in different colors.
Figure 4: Reconstructed dimuon invariant mass distribu- tion simulated for central Au+Au collisions with a beam momentum of 4 GeV/c (black histogram). The dimuon decays of low-mass vector mesons including their Dalitz decays, which have been identified using Monte Carlo par- ticle information, are shown in different colors.
Figure 3: Acceptance for omega mesons decaying in to muon pairs as function of transverse momentum and ra- pidity, for central Au+Au collisions at a beam momentum of 4 GeV/c. Black dots: reconstructed omega mesons for a muon detection system with 3 absorbers (60 cm C + 20+20 cm Fe). Colored area: 47 distribution generated by the PLUTO code. The black arrow indicates mid-rapidity. The signal-to-background ratios and the reconstruction efficiencies for omega mesons via their dimuon decay in central Au+Au collisions at beam momenta of 2 A GeV/c and 4 A GeV/c are listed in table 1.
Figure 3: Acceptance for omega mesons decaying in to muon pairs as function of transverse momentum and ra- pidity, for central Au+Au collisions at a beam momentum of 4 GeV/c. Black dots: reconstructed omega mesons for a muon detection system with 3 absorbers (60 cm C + 20+20 cm Fe). Colored area: 47 distribution generated by the PLUTO code. The black arrow indicates mid-rapidity. The signal-to-background ratios and the reconstruction efficiencies for omega mesons via their dimuon decay in central Au+Au collisions at beam momenta of 2 A GeV/c and 4 A GeV/c are listed in table 1.
Figure 5: Signal-to-background ratio as function of in- variant dimuon mass for reconstructed low-mass-vector mesons from central Au+Au collisions with 4 A GeV/c beam momentum.
Figure 5: Signal-to-background ratio as function of in- variant dimuon mass for reconstructed low-mass-vector mesons from central Au+Au collisions with 4 A GeV/c beam momentum.
Summary and conclusions
Summary and conclusions
Figure 1: Combinatorial unlike sign background from same event The invariant mass distributions are obtained by pair- ing different charge combinations muon candidates, like, ++, - - and +-.The combinatorial background is estimated from UrQMD events only, with same event (SE) and mixed event (ME) technique. In SE, pairs are constructed from unlike charge combinations muon candidates from same vent. Whereas, in ME, unlike pairs are formed from muon candidates always taken from different events . Fig. 1 and Fig. 2 shows the background distribution calculated from these two methods.
Figure 1: Combinatorial unlike sign background from same event The invariant mass distributions are obtained by pair- ing different charge combinations muon candidates, like, ++, - - and +-.The combinatorial background is estimated from UrQMD events only, with same event (SE) and mixed event (ME) technique. In SE, pairs are constructed from unlike charge combinations muon candidates from same vent. Whereas, in ME, unlike pairs are formed from muon candidates always taken from different events . Fig. 1 and Fig. 2 shows the background distribution calculated from these two methods.
Figure 4: Mixed event background superimposed with sig- nal w. As the signal peak over background is very small, so to get a significant yield of signal we need very high statistics for a stable background & ME background has been well suitable for its smooth shape due to large statistics. Here ME background has been scaled to the same lavel as same
Figure 4: Mixed event background superimposed with sig- nal w. As the signal peak over background is very small, so to get a significant yield of signal we need very high statistics for a stable background & ME background has been well suitable for its smooth shape due to large statistics. Here ME background has been scaled to the same lavel as same
Figure 3: background ratio of mixed event to same event Figure 2: Combinatorial unlike sign background from mixed event
Figure 3: background ratio of mixed event to same event Figure 2: Combinatorial unlike sign background from mixed event
event background. The ME background has been properly fitted with 5*” order polynomial. The total signal + ME background along with the background fit have been shown in Fig. 4. Finally fit is subtracted from the total S+B spec- tra to get the signal w yield as shown in Fig. 5. We get pair reconstruction efficiency of w ~0.97 % which is compara- ble with the signal efficiency obtained from true dimuons using PDG information.
event background. The ME background has been properly fitted with 5*” order polynomial. The total signal + ME background along with the background fit have been shown in Fig. 4. Finally fit is subtracted from the total S+B spec- tra to get the signal w yield as shown in Fig. 5. We get pair reconstruction efficiency of w ~0.97 % which is compara- ble with the signal efficiency obtained from true dimuons using PDG information.
Figure 2: 2D =~ reconstruction efficiency as a function of p, and rapidity (upper left), integrated =~ reconstruction efficiency as a function of p; (lower left) and rapidity (up- per right). Selected =~ invariant mass signal and total mass plots (lower right). Figure 1: Reconstructed invariant mass distribution of K°, A, A, =>, =t and Q7 in central Au+Au collisions at 10 AGeV, the red line indicates the signal plus background fit by a polynomial plus Gaussian functions.
Figure 2: 2D =~ reconstruction efficiency as a function of p, and rapidity (upper left), integrated =~ reconstruction efficiency as a function of p; (lower left) and rapidity (up- per right). Selected =~ invariant mass signal and total mass plots (lower right). Figure 1: Reconstructed invariant mass distribution of K°, A, A, =>, =t and Q7 in central Au+Au collisions at 10 AGeV, the red line indicates the signal plus background fit by a polynomial plus Gaussian functions.
I. Vassiliev', I. Kisel'?°, M. Zyzak', and the CBM Collaboration
I. Vassiliev', I. Kisel'?°, M. Zyzak', and the CBM Collaboration
Figure 1: The 2D plots for net-proton and net-kaon multiplicity distributions from reconstructed tracks (a), incident particles (b) and after unfolding (c) in Au-Au collisions at Ejqp = 10 AGeV.
Figure 1: The 2D plots for net-proton and net-kaon multiplicity distributions from reconstructed tracks (a), incident particles (b) and after unfolding (c) in Au-Au collisions at Ejqp = 10 AGeV.
Figure 2: The centrality dependence of Kok in Au-Au col- lisions at Ejay = 10 AGeV.
Figure 2: The centrality dependence of Kok in Au-Au col- lisions at Ejay = 10 AGeV.
where c is the velocity of light. Figure 1 shows m? dis- tribution of charged particles in the transverse momentum range 0.2 < pr < 2.0 GeV/c. Three well separated peaks are clearly visible which corresponds to 7, K and proton respectively. The protons and anti-protons are selected us- ing m? cut of 0.6 < m? < 1.2 GeV?/c* with purity of more than 96%. We have observed that number of anti-proton is very less compared to that of proton. Figure 1: Distribution of squared the mass (m7) of charged particles in transverse momentum range 0.2 < pr < 2.0 GeV/c. Vertical dotted lines indicate m? region for proton and antiproton selection.
where c is the velocity of light. Figure 1 shows m? dis- tribution of charged particles in the transverse momentum range 0.2 < pr < 2.0 GeV/c. Three well separated peaks are clearly visible which corresponds to 7, K and proton respectively. The protons and anti-protons are selected us- ing m? cut of 0.6 < m? < 1.2 GeV?/c* with purity of more than 96%. We have observed that number of anti-proton is very less compared to that of proton. Figure 1: Distribution of squared the mass (m7) of charged particles in transverse momentum range 0.2 < pr < 2.0 GeV/c. Vertical dotted lines indicate m? region for proton and antiproton selection.
tons and anti-protons are counted within the transverse mo- mentum range 0.2 < pr < 2.0 GeV/c and the rapidity range 1.2 < y < 2.2. The y — pr acceptance for pro- ton selection in Au-Au collisions at 10 AGeV is shown in Fig. 2. The reconstruction efficiency of protons and anti- protons in different centrality varries between 64 - 69 %. Figure 3 shows the event-by-event reconstructed (without efficiency correction) net-proton multiplicity distributions for different centrality bins. Figure 2: y—pr acceptance for protons in Au-Au collisions at 10 AGeV. The box shows the y — pr region used for the present analysis.
tons and anti-protons are counted within the transverse mo- mentum range 0.2 < pr < 2.0 GeV/c and the rapidity range 1.2 < y < 2.2. The y — pr acceptance for pro- ton selection in Au-Au collisions at 10 AGeV is shown in Fig. 2. The reconstruction efficiency of protons and anti- protons in different centrality varries between 64 - 69 %. Figure 3 shows the event-by-event reconstructed (without efficiency correction) net-proton multiplicity distributions for different centrality bins. Figure 2: y—pr acceptance for protons in Au-Au collisions at 10 AGeV. The box shows the y — pr region used for the present analysis.
Figure 3: The net-proton multiplicity distributions in dif- ferent centrality bins in Au-Au collisions at 10 AGeV.
Figure 3: The net-proton multiplicity distributions in dif- ferent centrality bins in Au-Au collisions at 10 AGeV.
Figure 4: The centrality dependence of cumulants in Au- Au collisions at 10 AGeV. The solid square markers show the centrality bin width corrected cumulants of recon- structed net-proton, open square markers show the cumu- lants of incident net-proton and solid traingle markers show the cumulants of unfolded net-proton distributions. The cu- mulants are shown upto 4*” order.
Figure 4: The centrality dependence of cumulants in Au- Au collisions at 10 AGeV. The solid square markers show the centrality bin width corrected cumulants of recon- structed net-proton, open square markers show the cumu- lants of incident net-proton and solid traingle markers show the cumulants of unfolded net-proton distributions. The cu- mulants are shown upto 4*” order.
Figure 1: Mean participant-zone rapidity shift ((yo)) as a function of Npart-
Figure 1: Mean participant-zone rapidity shift ((yo)) as a function of Npart-
'Department of Physics, University of North Bengal, Siliguri-734013, West Bengal, India; Department of Physics. Siliguri College, Siliguri-734001, West Bengal, India S. Sarkar‘, P. Mali’, and A. Mukhopadhyay!
'Department of Physics, University of North Bengal, Siliguri-734013, West Bengal, India; Department of Physics. Siliguri College, Siliguri-734001, West Bengal, India S. Sarkar‘, P. Mali’, and A. Mukhopadhyay!
Table 1: Fit parameter values corresponding to different <| yo |>.
Table 1: Fit parameter values corresponding to different <| yo |>.
Figure 3: Event by event participant-zone rapidity shift (yo) as a function of (a) spectator asymmetry (Aspect) and (b) participant asymmetry (Qpart) for the 20 — 25% cen- trality class.
Figure 3: Event by event participant-zone rapidity shift (yo) as a function of (a) spectator asymmetry (Aspect) and (b) participant asymmetry (Qpart) for the 20 — 25% cen- trality class.
Figure 4: (Color on line) Spectator asymmetry (Aspect) de- pendence of the mean participant-zone rapidity shift (yo) at different centrality classes. Qpart for the 20 — 25% centrality class is presented in Fig. 3. This unveils the exclusive correlation between yo and Qpart- The dispersion in eventwise yo values at a particular Qspec 18 also established which can be understood in terms of the term (A + B) present in Eq.(3). Fig. 4 shows the
Figure 4: (Color on line) Spectator asymmetry (Aspect) de- pendence of the mean participant-zone rapidity shift (yo) at different centrality classes. Qpart for the 20 — 25% centrality class is presented in Fig. 3. This unveils the exclusive correlation between yo and Qpart- The dispersion in eventwise yo values at a particular Qspec 18 also established which can be understood in terms of the term (A + B) present in Eq.(3). Fig. 4 shows the
Figure 5: (Color on line) Ratio of dN /dy for events with positive asymmetry (yo > 0) and those with negative asymmetry (yo < 0) as function of rapidity (y) for differ- ent centrality classes. The solid lines are fits to third order polynomial.
Figure 5: (Color on line) Ratio of dN /dy for events with positive asymmetry (yo > 0) and those with negative asymmetry (yo < 0) as function of rapidity (y) for differ- ent centrality classes. The solid lines are fits to third order polynomial.
Figure 1: (Color on line) Cumulant ratios of net-proton dis- tribution as a function of A7 in various pr windows. Department of Physics, University of North Bengal, Siliguri-734013, West Bengal, India S. Ghosh, P. Mali, and A. Mukhopadhyay the trivial volume dependence of the cumulants and are ex- pressed as,
Figure 1: (Color on line) Cumulant ratios of net-proton dis- tribution as a function of A7 in various pr windows. Department of Physics, University of North Bengal, Siliguri-734013, West Bengal, India S. Ghosh, P. Mali, and A. Mukhopadhyay the trivial volume dependence of the cumulants and are ex- pressed as,
Figure 3: (Color on line) Cumulant ratios of net charge distribution as a function of Av for various pr windows. Figure 2: (Color on line) Cumulant ratios of net kaon dis- tribution as a function of A7 for various pr windows.
Figure 3: (Color on line) Cumulant ratios of net charge distribution as a function of Av for various pr windows. Figure 2: (Color on line) Cumulant ratios of net kaon dis- tribution as a function of A7 for various pr windows.
Figure 1: The upper picture shows the engineering design of the mCBM setup inside cave-D (HTD). The middle and lower pictures sketch the data transport path, via the first FPGA layer (middle picture: AFCKs, MicroTCA crate) lo- cated inside the DAQ container, connected by optical fibers with the input as well as processing stage mounted inside the Green-IT-Cube (lower panels). 'GSI, Darmstadt, Germany; *Ruprecht-Karls-Universitat Heidelberg, Germany
Figure 1: The upper picture shows the engineering design of the mCBM setup inside cave-D (HTD). The middle and lower pictures sketch the data transport path, via the first FPGA layer (middle picture: AFCKs, MicroTCA crate) lo- cated inside the DAQ container, connected by optical fibers with the input as well as processing stage mounted inside the Green-IT-Cube (lower panels). 'GSI, Darmstadt, Germany; *Ruprecht-Karls-Universitat Heidelberg, Germany
Figure 2: Design of the new mCBM target chamber, man- ufactured at GSI.
Figure 2: Design of the new mCBM target chamber, man- ufactured at GSI.
Figure 3: Photograph of the new mCBM target chamber, equipped with a five-fold target ladder (right) and a 8x strip diamond TO counter (left). The width of the 8x vertical strips of the TO detector measures to 2mm each.
Figure 3: Photograph of the new mCBM target chamber, equipped with a five-fold target ladder (right) and a 8x strip diamond TO counter (left). The width of the 8x vertical strips of the TO detector measures to 2mm each.
Figure 4: Design of the 5-fold target ladder.
Figure 4: Design of the 5-fold target ladder.
Figure 6: Photography of the mCBM setup as of December 14, 2018 (beam: right to left). Figure 5: ROOT geometry of the mCBM test setup (ROOT geometry) at the HTD cave as of December 2018 (beam: left to right).
Figure 6: Photography of the mCBM setup as of December 14, 2018 (beam: right to left). Figure 5: ROOT geometry of the mCBM test setup (ROOT geometry) at the HTD cave as of December 2018 (beam: left to right).
Figure 7: Realized mCBM readout chain for the first commissioning campaign, based on DPB and FLIB. The mC BM subsystems are equipped with individual front-end electronics FEE (1). These front-ends are interfaced by the GB GB’ [x ASIC, which forwards the detector data via optical T link. All GBT links are received by the DBP layer lo- cated at 50 m distance in the DAQ container (2). The DPB isa FPGA based board which allows for subsystem specific pre- processing of the arriving data stream. A long distance optical link connects the DPB output to the FLIB board in- stall ed in the FLES input node in the Green IT Cube (3).
Figure 7: Realized mCBM readout chain for the first commissioning campaign, based on DPB and FLIB. The mC BM subsystems are equipped with individual front-end electronics FEE (1). These front-ends are interfaced by the GB GB’ [x ASIC, which forwards the detector data via optical T link. All GBT links are received by the DBP layer lo- cated at 50 m distance in the DAQ container (2). The DPB isa FPGA based board which allows for subsystem specific pre- processing of the arriving data stream. A long distance optical link connects the DPB output to the FLIB board in- stall ed in the FLES input node in the Green IT Cube (3).
Figure 8: Shift crew during data taking on December 14, 2018.
Figure 8: Shift crew during data taking on December 14, 2018.
Figure 1: Photograph of the HADES RICH photon detec- tor with most of the 428 Hamamatsu H12700 MAPMTs already mounted (Photo: G. Otto, GSD.
Figure 1: Photograph of the HADES RICH photon detec- tor with most of the 428 Hamamatsu H12700 MAPMTs already mounted (Photo: G. Otto, GSD.
The photon detector is operated in the fringe field region close to the HADES superconducting magnet. Cylindri- cal soft iron shields surrounding the PMT plane help to re- duce the magnetic field in the region of the MAPMTs to ac- ceptable values (2 mT at maximum magnet current). The magnetic field, as well as temperature of the radiator gas and backplanes, is measured using custom sensor boards mounted below- and close to the MAPMTs. Figure 2: Photograph of the electronic readout modules mounted backwards of the PMTs (Photo: G. Otto, GSI).
The photon detector is operated in the fringe field region close to the HADES superconducting magnet. Cylindri- cal soft iron shields surrounding the PMT plane help to re- duce the magnetic field in the region of the MAPMTs to ac- ceptable values (2 mT at maximum magnet current). The magnetic field, as well as temperature of the radiator gas and backplanes, is measured using custom sensor boards mounted below- and close to the MAPMTs. Figure 2: Photograph of the electronic readout modules mounted backwards of the PMTs (Photo: G. Otto, GSI).
Figure 3: Close view of the LV supply arms and copper grounding ring, and of the mounting structures for individ- ual readout modules.
Figure 3: Close view of the LV supply arms and copper grounding ring, and of the mounting structures for individ- ual readout modules.
Figure 4: Hit rates (in Hz) on each of the 27k readout Pixels without beam (left) and with beam on target (right). Green: ~20 Hz, red: >100 kHz
Figure 4: Hit rates (in Hz) on each of the 27k readout Pixels without beam (left) and with beam on target (right). Green: ~20 Hz, red: >100 kHz
Figure 5: Leading edge time distribution with respect to the trigger for data taken with the CaF2 window still in place.
Figure 5: Leading edge time distribution with respect to the trigger for data taken with the CaF2 window still in place.
Figure 6: View of the open HADES RICH radiator ves- sel with the multi-hexagon-segmented CaF) window still mounted in front of the focusing mirror.
Figure 6: View of the open HADES RICH radiator ves- sel with the multi-hexagon-segmented CaF) window still mounted in front of the focusing mirror.
Dr Valentina Akishina, winner of the CBM PhD Award 2017
Dr Valentina Akishina, winner of the CBM PhD Award 2017

Key takeaways
sparkles

AI

  1. The CBM experiment aims to explore the QCD phase diagram at high baryon densities with unprecedented data rates.
  2. CBM is transitioning from R&D to large-scale production and integration of detector hardware by 2024.
  3. The full-system test setup mCBM allows joint operation of multiple detector systems for data acquisition.
  4. Unique measurements like rare hadron production and di-electron distributions require robust detector systems and software.
  5. The report outlines ongoing advancements in detector technology, calibration, and the integration of electronics for high-rate environments.

Related papers

CBM - Compressed Baryonic Matter experiment at FAIR - Progress Report 2014
Wojciech Zabołotny

2015

Dear colleagues, by now, the CBM Progress Report already looks back at a considerable history. Its first issue dates back to the year 2006, when we decided to collect all CBM-related contributions to the GSI Scientific Report, which at that time ceased to appear in print, in a separate, printed volume, and augment it with numerous CBM reports from outside GSI. Since then, the CBM Progress Report annually documents the manifold activities towards the realisation of the CBM project, and its ever increasing size demonstrates the steady growth of efforts. Browsing through the past volumes, it is possible to trace some of the CBM history, from first conceptual considerations and rather basic feasibility studies, over thorough detector and electronics R&D to a close-tofinal detector design and physics performance studies with detailed and realistic detector response. Now, the year 2014 marks the beginning of a new phase of the CBM project. Most of the Technical Design Reports are submitted and approved, the basic technological and design decisions are made, and the physics programme is being shaped for the first years of operation at the SIS-100 accelerator. The next years will be devoted to the building of the detectors, the actual realisation of what has been developed and planned over the past years. The FAIR project itself enters a new phase. The managements of both FAIR and GSI are being restructured, and construction plans, priorities and time lines are under renewed discussions. Whatever the outcome will be, CBM continues to be one of the most important scientific pillars of the FAIR, and our efforts remain directed to have our experiment ready for taking data when the FAIR accelerators will start to deliver beams. Our thanks go to all who contributed to this report.

View PDFchevron_right
CBM Experiment Local and Global Implications
Wojciech M Zabolotny, Ryszard Romaniuk

—The research area of the compressed baryonic matter-CBM experiment (FAIR/GSI in Darmstadt) is sub-nuclear physics, thus hadron-baryon and quark-gluon, and the essence of phase transitions in the area of hot nuclear matter, and dense strongly interacting matter. Our interest in this paper are mainly considerations on the impact of such large infrastructural experiments and possibilities they give to local, smaller but very active, university based research groups and communities. Research and technical input from such groups is depicted on the background of the CBM detector infrastructure and electronic instrumentation just under design and test fabrication for this experiment. An essential input to this research originates from Poland via the agreed in-kind contribution. The areas of expertise of these groups are: superconductivity, structural large scale cabling, precision machined parts, RF and microwave technology, analog and advanced digital electronics, distributed measurement and control systems, etc. Keywords—CBM experiment, compressed baryonic matter, advanced electronic systems, measurement systems, DAQ systems, FAIR, GSI, European large research infrastructures

View PDFchevron_right
Compressed Baryonic Experiment at FAIR; CBM Progress Report 2017
Ryszard Romaniuk

Alberica Toia, Ilya Selyuzhenkov, (Edit.), Compressed Baryonic Experiment at FAIR; CBM Progress Report 2017, 2018

Editors: Ilya Selyuzhenkov (ilya.selyuzhenkov@gmail.com), Alberica Toia (a.toia@gsi.de); Reviewers: P.P. Bhaduri, C. Blume, S. Chattopadhyay, J. de Cuveland, I. Deppner, D. Emschermann, V. Friese, M. Golubeva, F. Guber, N. Herrmann, J. Heuser, C. Höhne, W. Müller, C. Müntz, I. Selyuzhenkov, P. Senger, C. Sturm, A. Toia; Contributors: Members of CBM Collaboration; DOI: 10.15120/GSI-2018-00485 (https://dx.doi.org/10.15120/GSI-2018-00485); ISBN 978-3-9815227-5-4; Printed in Darmstadt by GSI, March 2018;

View PDFchevron_right
Astrophysics in the Laboratory—The CBM Experiment at FAIR
P. Senger

Particles, 2020

The future “Facility for Antiproton and Ion Research” (FAIR) is an accelerator-based international center for fundamental and applied research, which presently is under construction in Darmstadt, Germany. An important part of the program is devoted to questions related to astrophysics, including the origin of elements in the universe and the properties of strongly interacting matter under extreme conditions, which are relevant for our understanding of the structure of neutron stars and the dynamics of supernova explosions and neutron star mergers. The Compressed Baryonic Matter (CBM) experiment at FAIR is designed to measure promising observables in high-energy heavy-ion collisions, which are expected to be sensitive to the high-density equation-of-state (EOS) of nuclear matter and to new phases of Quantum Chromo Dynamics (QCD) matter at high densities. The CBM physics program, the relevant observables and the experimental setup will be discussed.

View PDFchevron_right
Preface: physica status solidi (c) 7/2015
Patrik Hoffmann

physica status solidi (c), 2015

View PDFchevron_right
CERN summer school report (2010)
Mohammed El Amine Benaichouche

2010

One week after obtaining my master degree in Theoretical Physics, there is no place better than CERN to start learning the techniques and the methods used to compare the theoretical predictions of with experimental results. I believe that the work in experimental particle physics is really different from the work on the mathematical physics. In the discussions with my supervisors Prof. Krzysztof Piotrzkowski and Dr. Jonathan Hollar, we were trying to find a good project and the best path to go from Theoretical physics to Particle physics, because I decided to start my career and do my PhD in experimental high energy physics. I think that the modern theory beyond the standard model are very far from the experimental evidence now and the work on confronting these theories with data or on searching for new physics phenomena is more attractive to me.

View PDFchevron_right
P. Biagi, E. Starnini, B.A. Voytek 2018 - Summary
Barbara Voytek, Elisabetta Starnini, Paolo Biagi

Paolo Biagi e Elisabetta Starnini (a cura di) GLI SCAVI ALL’ARMA DELL’AQUILA (FINALE LIGURE, SAVONA): LE RICERCHE E I MATERIALI DEGLI SCAVI DEL NOVECENTO. Quaderni della Società per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia, 15., 2018

Summary of the volume

View PDFchevron_right
Compressed Baryonic Matter experiment at FAIR, Progress Report 2022
Ryszard Romaniuk

CBM Progress Report 2022, 2023

Compressed Baryonic Matter experiment at FAIR; CBM Progress Report 2022, CBM Collaroration; Editor Peter Senger, Co-editor Volker Friese; Reviewers: C. Blume, S. Chattopadhyay, J. de Cuveland, I. Deppner, D. Emschermann, V. Friese, J. Heuser, C. Höhne, M. Kis, C. Müntz, I. Selyuzhenkov, P. Senger, C. Sturm, A. Toia; Published 2022 by GSI Helmholtzzentrum für Schwerionenforschung GmbH, Darmstadt, Germany

View PDFchevron_right
Compressed Baryonic Matter experiment at FAIR, Progress Report 2019
Ryszard Romaniuk, F. Volker

CBM Progress Report 2019, 2020

The exploration of the QCD phase diagram at high net-baryon densities is one of the most fascinating aspects of modern nuclear physics, as it links laboratory heavy-ion collisions to cosmic objects and phenomena like neutron stars, supernova explosions, and neutron star mergers. Moreover, accelerator based experiments have the potential to unravel the microscopic degrees-of-freedom of strongly interaction matter at high density, which may undergo phase transitions, featuring phase coexistence and a critical endpoint. The mission of the CBM experiment is to shed light on the nature of high-density QCD matter by investigating diagnostic probes, which never have been measured before in the FAIR energy range, where the highest net-baryon densities will be created.

View PDFchevron_right
Scientific Event: Isreie 2019
Dana Copolovici

2019

Within the general program, 10 parallel sessions were carried out. In the present article, we have been focused on 3 sessions that were organized by the

View PDFchevron_right

Related papers

CERN openlab annual report 2019
Federico Carminati

2020

View PDFchevron_right
Suba, E., Szalma, E., Szucs, A., Schön, S., Cullen, J., Fischer, T., Kuiper, E.R. & Hartog, M. (2012). Production and Dissemination: Final Report Events
M. Hartog

2012

View PDFchevron_right
D1.2 First interim scientific report
Sergio Pareja ruiz

2013

View PDFchevron_right
CHIPP Roadmap for Research and Infrastructure 2025–2028 and beyond by the Swiss Particle Physics Community
34.Thùy Nguyễn

2021

View PDFchevron_right
Science with the New Generation of High Energy Gamma-Ray Experiments
SUSHIL K VERMA

2007

View PDFchevron_right
Science with the new generation high energy gamma-ray experiments
Miguel Carvajal

2007

View PDFchevron_right
Annual Report of the Helmholtz-Institute Jena 2018
K Schmal

2017

View PDFchevron_right
CERN openlab Whitepaper on Future IT Challenges in Scientific Research
Mélissa Gaillard

2014

View PDFchevron_right
BG-2020-126 Review
Tamar Guy-Haim

2020

View PDFchevron_right
Horizons Review Symposium
Charles Camosy

Horizons, 2017

View PDFchevron_right
Working Group Report: Physics
B. Brahmachari

2004

View PDFchevron_right
The Compressed Baryonic Matter Experiment at FAIR
Krzysztof Pozniak, Ryszard Romaniuk

Nuclear Physics A, 2013

View PDFchevron_right
Future Experiments : Project X
Sanjib Mishra

Proceedings of 10th International Workshop on Neutrino Factories, Super beams and Beta beams — PoS(Nufact08), 2009

View PDFchevron_right
Software Development for the BM@N Experiment at NICA: Challenges and Status
Konstantin Gertsenberger

EPJ Web of Conferences, 2020

View PDFchevron_right
Institute of Nuclear Physics Polish Academy of Sciences Report on research
Wojciech Kwiatek

2014

View PDFchevron_right
Report of the First Fusion Energy Sciences Committee of Visitors
Michael E Mauel

Journal of Fusion Energy, 2003

View PDFchevron_right
Final Summary Report
Chris Cocklin

Deliverable D5, 2006

View PDFchevron_right

Related topics

PhysicsHigh Energy PhysicsParticle AcceleratorsCompressed Baryonic Matter
580 California St., Suite 400
San Francisco, CA, 94104