David Berge Raffaele Giordano George Glonti George Iakovidis

+ David Berge, Raffaele Giordano, George Glonti, George Iakovidis, Paolo Iengo, Vincenzo Izzo, Kostas Kordas, Antonios Leisos, Sabrina Perrella, Givi Sekhniaidze, Ourania Sidiropoulou, Mimmo della Volpe, Andre Zibell SRS development: Hans Muller, Sorin Martoiu, Alfonso Martinez (ALICE) Marcin Byszewski On behalf of the MAMMA collaboration. Micromegas in ATLAS: status update of DAQ integration and plans 25/09/2012 MAMMA extended M. Byszewski(CERN)

+ 2 The Goal Compare data from the test chambers with ATLAS data (take these data with no impact on ATLAS data taking) Stand-alone, random trigger (until September 2012) 1. n Convenient, Track matching not possible Trigger from ATLAS, data separate (‘parasitic’) 2. n Offline synchronisation Fully integrated with TDAQ in ATLAS partition 3. n Most of our event fragments empty A lot of discussions and support from CTP / TDAQ / run coordination / Sysadmins. Thank you. M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 3 Micromegas test chambers Installed Feb 2012, read out in stand-alone random trigger mode Test chambers description and readout in stand alone mode see Joerg Wotschack’s presentation in Run Weekly 24/7/2012 https: //indico. cern. ch/conference. Display. py? conf. Id=194946 MBTS, side A n n Small Wheel, sec. 9, side A, CSC Front of the LAr calorimeter cryo n r≈1 m n z = 3. 5 m One 9 x 4. 5 cm 2 n X-V (2 readout gas gaps) n n r = 1. 7 -1. 8 m Four 9 x 9 cm 2 chambers n X, Y, XUV n Only three read out n 1 FEC limit n Power supply limits and SRU firmware LAr EIL 1 MDTs MBTS M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 4 Rates / Occupancy / data size MBT 0 n n Rates: n ≈20 k. Hz/cm 2 @ L=1033 cm-2 s-1 n 7 strips (16 time bins) 10 k. Hz readout n (100 k. Hz * 0. 1 due to slow readout) n Up to 10 MB/s by varying time window and data reduction mode (36 GB/h) n Total data: for 10 weeks, 5 days, 10 h runs 17 TB of LV 1 data M. Byszewski(CERN) SW n Rate: n 30 Hz/cm 2 @ L = 1033 cm-2 s- n Majority (90%) of events with uncorrelated hits 1 25/09/2012 MAMMA extended

+ 5 Current readout: ½ ROD n SRS – based ROD: n n RD 51’s SRS system https: //espace. cern. ch/rd 51 -wg 5/srs/default. aspx n APV 25 chips (CMS Si tracker) (16+4) n HDMI cables (10) n SRS FEC (ADC, Ethernet) Data to a DAQ PC in USA 15 UX 15 APV 25 HDMI M. Byszewski(CERN) FEC (ADC/Eth ernet) Ethernet fibre Slow control + Data Switch (Cu Optical) Cu DAQ PC 25/09/2012 MAMMA extended

+ 6 SRS – based Readout n n n Front end electronics: n APV 25 chips (CMS tracker, no other choice) n HDMI cables (LV, data) ROD in UX 15: n SRS FEC – digitization, peak finding, zero suppression n DTC link to SRU n SRU – EB, TTC, LV 1, DCS , SLINK USA 15 n CSC TTC, DATA, DTC fibres n Run Control Application (on RC PC) n ROS n DCS(ACR Muon Desk, CSC infrastructure) USA 15 Ctrl UX 15 ROD APV 25 FEC (ADC/Eth ernet) Switch (Cu Optical) SRU RC PC ROS Local SLINK TTC CSC M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 7 Configuration (1): TDAQ 1) Fully integrated solution required much more work (MM and TDAQ) without clear advantage for data analysis. 2) Parasitic mode n n n ROS LVL 1 A from CSC TTC crate Sub-detector ID : RPC 0 x 65 (side A), data channel 0 x. FF (nonexistent, ignored by decoder) Separate MM partition for ROS local storage RC PC n Offline synchronization with ATLAS data ATL RC Parasitic with RCD in ATLAS partition 3) n n Send UDP packages to our ROD (e. g. , to set Run. Number) (Always returns with success) M. Byszewski(CERN) Local 25/09/2012 MAMMA extended

+ 8 Configuration (2): parasitic n We will have all the data we measure n There is no way of getting LV 2/EF information to our ROS (event selection) n Must store LV 1 data as long as needed to get ATLAS accepted LV 1 n 1 TB storage for temporary data n (all our LV 1, a few days of data taking) M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 9 Configuration (3): ROD n Not in ATLAS partition n n MM partition fully described in TDAQ OKS database Answer every LV 1 trigger Reading out 10% of LV 1 triggers (slow APV data transfer) Pre-selection on SRU possible n Send only events with APV data n Possibly select on FEC event size We could switch back n n into fully integrated mode and serve all LV 1 if in ATLAS partition ROD (1) APV 25 M. Byszewski(CERN) FEC (ADC/Eth ernet) SRU ROS 25/09/2012 MAMMA extended

+ 10 Run Modes n n n Castor ATL CTRL net Physics n ATLAS LV 1, 10% n Data to SLink-ROS-RC PC-HDD n Offline synchronisation of LV 1 n Throttled data transfer to storage (Castor) n All APV / SW-only runs l ro nt w o Sl Off-run n Internal / CSC triggers n On FEC measurement of ZS pedestals n RC PC storage of pedestals (read from FEC) APV calibration (rare) n Internal triggers n ZS in bypass mode n SRU in bypass mode to Ethernet port n Display / Verify raw APV frames M. Byszewski(CERN) co Local ETH Calibration n Throttled ctrl RC Eth RO S SLINK CSC TTC SRU FEC 25/09/2012 MAMMA extended

+ 11 Run Modes n n n Castor ATL CTRL net Physics n ATLAS LV 1, 10% n Data to SLink-ROS-RC PC-HDD n Offline synchronisation of LV 1 n Throttled data transfer to storage (Castor) n All APV / SW-only runs l ro nt w o Sl Off-run n Internal / CSC triggers n On FEC measurement of ZS pedestals n RC PC storage of pedestals (read from FEC) APV calibration (rare) n Internal triggers n ZS in bypass mode n SRU in bypass mode to Ethernet port n Display / Verify raw APV frames M. Byszewski(CERN) co Local ETH Calibration n Throttled ctrl RC Eth RO S SLINK CSC TTC SRU FEC 25/09/2012 MAMMA extended

+ 12 Run Modes n n n Castor ATL CTRL net Physics n ATLAS LV 1, 10% n Data to SLink-ROS-RC PC-HDD n Offline synchronisation of LV 1 n Throttled data transfer to storage (Castor) n All APV / SW-only runs l ro nt w o Sl Off-run n Internal / CSC triggers n On FEC measurement of ZS pedestals n RC PC storage of pedestals (read from FEC) APV calibration (rare runs) n Internal triggers n FEC ZS in bypass mode (RAW data of one APV) n SRU in bypass mode (to Ethernet port) n Display / Verify raw APV frames M. Byszewski(CERN) co Local ETH Calibration n Throttled ctrl RC Eth RO S SLINK CSC TTC SRU FEC 25/09/2012 MAMMA extended

+ 13 Offline synchronisation Castor ATL CTRL net n Before data analysis n Save all RAW events n n n n Throttled ATLAS synchronisation at our convenience Max of 16 TB of data (SW+MBTS) for 10 weeks of 10 h ATLAS runs daily We will get that space Throttled speed of output to storage will limit data taking Unless: Dedicated SW runs Automatic RC Synchronise “on the fly” n n Extract LV 1 from accepted ATLAS events (delay of few days, ATHENA) Select matching events and delete other Run storage job Limits: Storage buffer size, disk r/w speed, CPU speed, output speed, limited access to P 1 network M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 14 Current status n Readout synchronised with ATLAS triggers n n 1 FEC (MBT 0 + SW chambers) n n There is no way of getting LV 2/EF information from TDAQ Power supply RC-PC with storage space is being prepared n Time to debug M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 15 Next steps: to data analysis n Run planning to be made (10 weeks) n n n DAQ commissioning n n n MBT 0/SW runs dedicated SW runs (re-cabling 4 th chamber ? ) Verify APV settings (raw data) Verify FEC ZS pedestals Stability (not to disturb the Muon shifter’s sleep) Write data decoder (ATLAS -> ROOT for Event. Browser) Data preparation / Offline Synchronisation n Select based on Timestamp + LV 1 ID (ATLAS and MM) Select events with CSC tracks ATHENA jobs n Data analysis n (Write Manual / Documentation) M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 16 Thank you n n Call for expression of interest n DAQ verification / debugging n Data preparation n Data analysis Continuing interest : n Andre, Antonis, Kostas, Ourania M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ 17 Backup slides M. Byszewski(CERN) 25/09/2012 MAMMA extended

Micromegas principle § With drift velocities of 5 cm/µs (or 20 ns/mm) electrons need 100 ns drift time to reach the mesh (for a 5 mm gap) § By measuring the arrival time of the signals a MM functions like a TPC => Track vectors for inclined tracks 25/09/2012 MAMMA extended 128 µm Conversion & drift region (typically a few mm) with moderate electric field of 100– 1000 V/cm Amplification in a narrow (128 µm) gap with high electrical field (40– 50 k. V/cm) 5 mm § Parallel-plate chamber M. Byszewski(CERN) 18

+ 19 SRU n n n Virtex 6, TTCrx chip, 4 SFP ports 40 DTC links EB, and TTC LVL 1 Accept treatment n n TTC n n Process first, buffer others SRU uses onboard TTCrx chip to receive BC clock, L 1 A, ECR, BCR and trigger type Connect to CSC TTC partition with unique TTC address Additional user-programmable offset value for BCID S-LINK n HOLA emulator on Virtex 6 board M. Byszewski(CERN) 25/09/2012 MAMMA extended

+ Some typical events displays 20 L = 3. 3 x 1030 cm-2 s-1 L = 3. 3 x 1033 cm-2 s-1 (≈1/1000 triggers with activity) (≈each trigger with activity) Time (25 ns) Charge Time (25 ns) MBT 0_3 Uncorrelated hit x strip number M. Byszewski(CERN) Time (25 ns) Charge Time (25 ns) MBT 0_4 x strip number 25/09/2012 MAMMA extended

150 mm MBT chambers (LAr ecal) MBTS R ≈ 1 m APV 25 Drift electrode 200 mm MBT 0 2 drift gaps of 4. 5 mm each 2 x 190 x(phi)-strips (0. 5 mm pitch) 2 x 66 v-strips (1. 5 mm pitch) MBT 0_3 Active area 9. 5 x 4. 5 cm 2 LAr Calor. Z ≈ 3. 5 m MBT 0_4 15 mm 25/09/2012 MAMMA extended M. Byszewski(CERN) 21

MMs on Small Wheel II R 13 x R 16 xy R 16 Not yet read out R 19 xuv R 18 y CSC (large) 90 mm Pitch (µm) R 13–R 18 250 R 19 x 350 R 19 u, v 100 25/09/2012 MAMMA extended R 16, chamber with 2 D readout active area: 9 x 9 cm 2 M. Byszewski(CERN) 22

+ M. Byszewski(CERN) 23 25/09/2012 MAMMA extended

+ M. Byszewski(CERN) 24 25/09/2012 MAMMA extended