1. New Small Wheel muon trigger optical module New Small Wheel muon trigger optical module S. Hou 2014/08/29 Academia Sinica

2. 2 2 Muon System p T measurement : Monitored Drift Tube (MDT) Cathode Strip Chamber (CSC) Measure deflection by toroidal field Barrel Trigger : Resistive Plate Chambers (RPC) Endcatp Trigger: Thin Gap Chambers (TGC)

3. 3 Challenges to ATLAS Phase-I upgrade LHC shutdown 14 months in 2018  Consolidation of injector chain, collimators  Peak luminosity 3 × 10 34 cm -2 s -1 Challenge to ATLAS  TRIGGER RATE OVERBOUND Event pileup up to 80 per bunch crossing  keep trigger threshold around 20-25 GeV  muon p T thresholds not effective in the forward region  higher EM E T thresholds in physics acceptance

4. 4 New Small Wheel upgrade Multilayer chambers Kill fake triggers at Level-1 by IP pointing δθ ~ 1 mrad  a trigger rate reduction ~6

5. 5 New Small Wheel detectors 16 sectors per wheel, each sector has 2x4 sTGC (Thin Gap Chambers) layers 2x4 MM (MicroMegas) layers sTGC MicroMegas

6. 6 New Small Wheel trigger upgrade add NSW sTGC trigger muon track δθ < 1 mrad reduce noise tracks ATLAS trigger limits L1 : ~75 kHz max L2 : ~3.5 kHz max EF : ~200 Hz max Forward muon trigger (kHz) at L= 3×10 34 MU11 events selected analysis Upgrade L1 with NSW

7. 7 NSW sTGC trigger circuits

8. 8 NSW sTGC router module Repeater to clean and amplify signals from TDS FPGA selects active TDS signals to transmit Multi-rate transceiver to bump the signal speed to ~10 Gbps Optical Tx send out over fiber

9. 9 sTGC router optical transceiver 1. Use VCSEL in TOSA 850 nm, MM, 10 GHz  no fiber alignment issue, TOSA takes LC fiber connecter production is plainly PCB SMD process 2. Choice of GBLD or LOCld drivers 3. Joint Optical project of Phase-I LAr+NSW ? Requires approval  LAr MTX is rad-hard, up to 8 Gbps (Xcheck?)  LAr requires 5000 lines  sTGC router is satisfied? Total 800 lines  only QA, no R&D required Test fanout board, SMA to I/O e.g. Kintex7, Scope SMU MTX module

10. 10 Optical transceiver production, QC/QA Manufacturers in contact 1. Liverag.com.tw Expertise on active optical products Capable of >10Gbps modules, QA, circuits 2. FOCI.com.tw Expertise on passive fiber assemblies Jointly with SMU, a visit to manufacturers is planned in early October  seeking all kind of collaboration opportunities and cost estimation Laboratory setup Bench test, scope for 10 GHz eye diagram Bit Error Rate using Kintex7 board Radiation tests, Co 60 gamma, proton 30 MeV

11. 11 QA on VCSELs: Samples −VCSELs of various manufacturers Die/Wire bonds by FOCI robot VCSEL dies on an 10×10 cm 2 PCB −TOSA of TrueLight assembled on PCB, no latch assembly

12. 12 DC light power measurements −V-I-L scan by a LabView setup VCSELs covered by a large (10 × 10 mm 2 ) GaAs PIN, Mechanical alignment is required NI 6024E PCMCIA to an XP notebook −Power meter measurement

13. 13 DC bias to VCSELS −DC bias by Agilent E3631A, Keithley 2304A current kept at ~ 5 mA/ch −Bias at 1.65 V, 1.75 V, 2.0 V, to VCSELs channel current measured, bundled.

14. 14 85/85 chamber −Temperature stable at 0.1 o C −Humidity stable at 0.2% RH −Cooled to 30/55 before opening, to prevent condensation Burn periods conducted 30/50 : 118 hrs 85/85 : 12 hrs 85/85 : 94 hrs 85/85 : 275 hrs 85/85 : 316 hrs 85/85 : 363 hrs Continuing..

15. 15 Reference samples −Not in oven, to examine the systematics of DAQ −Join afterwards, DAQ with test samples 1F45 1F59 1F58

16. 16 Reference samples −FINISAR 2092-001 5 Gb −board 2, two 4x1 arrays VCSELS are not centered at PIN Mechanical alignment is an issue

17. 17 TOSA samples are robust in 85/85 TrueLight TOSA 1F45 4.25 Gb 1F58 10 Gb 1F59 10 Gb No obvious loss After ~700 hr 85/85 Bad electric contact Boards were poorly prepared, ch8 light recovered 1F45 1F58 1F59

18. 18 FINISAR array degradation Only 1 of 8, showing large degradation, the rest are consistent with small degradation, in 1100 hrs @ 85/85