Presentation is loading. Please wait.

Presentation is loading. Please wait.

Algorithms for the ROD DSP of the ATLAS Hadronic Tile Calorimeter

Similar presentations


Presentation on theme: "Algorithms for the ROD DSP of the ATLAS Hadronic Tile Calorimeter"— Presentation transcript:

1 Algorithms for the ROD DSP of the ATLAS Hadronic Tile Calorimeter
Belén Salvachúa and Arantxa Ruiz-Martínez IFIC – Universidad de Valencia 12th Workshop on Electronics for LHC and Future Experiments 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

2 ATLAS detector 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

3 Hadronic Tile Calorimeter
12 m EBA LBA LBC EBC 4 m 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

4 Hadronic Tile Calorimeter
Los módulos del TileCal tienen una geometría proyectiva en eta como muestra la figura y están divididos en 3 capas: A, BC y D, cuya segmentación en eta es 0.1 para las dos más internas y 0.2 para la más externa. La energía depositada en cada celda es leída por 2 fotomultiplicadores. Con lo cual, TileCal dispone de un total de 4672 celdas y 9344 canales de lectura. 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

5 Read Out Chain SHAPER TILECAL MODULE PMT READ OUT DRIVER SYSTEM
DIGITIZER Optical Fibers 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

6 Read Out Driver board G-links 12th LECC06 - Valencia 28/09/2006
B. Salvachúa

7 Read Out Driver board Staging FPGAs 12th LECC06 - Valencia 28/09/2006
B. Salvachúa

8 Read Out Driver board Processing Units
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

9 Read Out Driver board Output Controller FPGAs
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

10 Read Out Driver board Serializers 12th LECC06 - Valencia 28/09/2006
B. Salvachúa

11 Read Out Driver board VME and TTC FPGA
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

12 Read Out Driver board Staging FPGAs Processing Units G-links
VME and TTC FPGA Optical Fibers Output Controller FPGAs Serializers Transition Module 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

13 Processing Units: DSP Eight functional units: 8/16/32-bit data support
2 multipliers 6 arithmetic and logical units 8/16/32-bit data support 40-bit arithmetic options Clock cycle of 720 MHz Memory: 1056 Kbytes 32 Kbytes cache 1024 Kbytes RAM Real time fixed-point processor TMS360C6414xTM Texas Instruments 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

14 Processing Units: DSP Eight functional units: 8/16/32-bit data support
2 multipliers 6 arithmetic and logical units 8/16/32-bit data support 40-bit arithmetic options Clock cycle of 720 MHz Memory: 1056 Kbytes 32 Kbytes cache 1024 Kbytes RAM Real time fixed-point processor TMS360C6414xTM Texas Instruments 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

15 TileCal commissioning setup
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

16 TileCal commissioning read out
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

17 Reconstruction Algorithms
Requirements: Send reconstructed information to the 2nd level trigger Work in real-time at 1st level trigger rate LHC rate: 100 kHz First years rate: ~50 kHz Commissioning rate (during July-August 2006): ~1Hz Proposed algorithms: Optimal Filtering: Reconstruction of the energy and arrival time of the particles Muon Tag: Identification of low transverse momentum muons 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

18 Optimal Filtering AMPLITUDE PHASE WEIGHTS The process to calculate a, b, minimizes the effect of the noise in the amplitude and time reconstruction. But they are calculated assuming small phases. 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

19 Optimal Filtering Configuration for LHC
AMPLITUDE PHASE WEIGHTS The process to calculate a, b, minimizes the effect of the noise in the amplitude and time reconstruction. But they are calculated assuming small phases. 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

20 Optimal Filtering with 1 iteration
Commissioning configuration First iteration: Second Iteration: AMPLITUDE AMPLITUDE PHASE PHASE 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

21 Optimal Filtering with 1 iteration
Commissioning configuration First iteration: Second Iteration: AMPLITUDE AMPLITUDE PHASE PHASE 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

22 Muon Tag Look into cells D  Possible muon
Look into cells BC, following   Possible muon Look into cells A, following   Muon identified Deposited energy in the cell i should verify: Thrlow < Ei < Thrhigh High threshold: Cut for jets Low threshold: Cut for noise and minimum bias El funcionamiento del algoritmo es el siguiente: en primer lugar se busca en la capa D las celdas con deposición de energía compatible con un muón, definiendo los primeros candidatos a muones. Tras esto se busca en la celdas de la capa BC situadas debajo de los candidatos definidos anteriormente, y si la energía de estas celdas es compatible también con un muón se confirma este candidato. Finalmente se mira si en las celdas de la capa A situadas en la misma eta que los candidatos la energía depositada también es la propia de un muón, confirmándose de este modo la identificación del muón. La condición que se aplica a todas las celdas es que la energía depositada este comprendida entre un umbral inferior y un umbral superior. El umbral inferior es utilizado para cortar el ruido electrónico y el apilamiento de minimum bias y se ha utilizado un valor de 150 MeV. El umbral superior depende de cada celda y descarta las cascadas hadrónica y sus colas. Esta condición es exigida en las celdas de las tres capas, pero los muones también son identificados si la energía es mayor que el umbral mínimo en todas las capas y menor que el máximo solo en 2, para permitir la identificación de muones que pierdan una parte apreciable de energía en una de las capas. 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

23 Results: Optimal Filtering
Amplitude (energy) reconstruction of cosmic muons For E > 3 of the noise differences < 1% 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

24 Results: Optimal Filtering
Phase interval [-15,15] ns Phase differences ~5% 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

25 Results: Muon Tag 85% of offline tagged muons are tagged online
12th LECC06 - Valencia 28/09/2006 B. Salvachúa

26 Result: Processing Time
Algorithm Time (s) Optimal Filtering 27.1 Optimal Filtering and Muon Tag 29.4 12th LECC06 - Valencia 28/09/2006 B. Salvachúa

27 Conclusions Optimal Filtering and Muon Tag algorithms were running during TileCal commissioning July and August 2006 Optimal Filtering: Amplitude accuracy > 99% for E>3 Phase accuracy around 95% for E>3 Muon Tag: 85% of coincidence offline/online Processing time: Opt. Filt and Muon Tag: 29.4 s Fulfills TileCal commissioning requirements Improvements on the time are expected in the change to assembler 12th LECC06 - Valencia 28/09/2006 B. Salvachúa


Download ppt "Algorithms for the ROD DSP of the ATLAS Hadronic Tile Calorimeter"

Similar presentations


Ads by Google