WBS 1.04 KLM Systems (Technical) Leo Piilonen Virginia Tech

Talk Outline WBS Element Scope Performance Requirements Technical Material Summary January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)2

WBS 1.04 Scope Design, fabricate, deliver to KEK and test the barrel KLM scintillator readout modules for layers 0 and 1. Specify, procure, implement and commission the high voltage power supplies for the barrel and endcap KLM scintillator modules. Design, implement and commission the KLM track trigger Implement the KLM-associated software January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)3

Performance Requirements (1) Detect K L mesons with high efficiency & purity and with good angular resolution – For CP-sensitive B decay modes like – For K L veto for missing energy modes like 2o2o Efficiency Angular resolution January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)4

Performance Requirements (2) Identify muons with high efficiency & purity for momenta above 0.6 GeV/c Efficiency Impurity January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)5

Performance Requirements (3) Efficiency & purity not degraded by ≈100 Hz/cm 2 ambient neutron background, unlike RPCs Scintillators match the requirements RPC efficiency falls at –0.05 per Hz/cm 2 Barrel RPC Efficiency January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)6

Neutron mitigation: Replace layer 0 and 1 RPCs with extruded-strip scintillator detectors within the same form-factor modules as the existing RPCs End view Scintillator or RPC module January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)7

Detect WLS-fiber light with Geiger-mode avalanche photodiode (“SiPM” or “MPPC”) 1.3 x 1.3 mm pixels Rubber spring Scintillator strip developed for T2K near detector operates in 1.5 T magnetic field rad-hard (>10-year lifetime in KLM) 8-pixel threshold gives >99% efficiency January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)8 Wavelength-shifting fiber Hamamatsu S10362 attached to one end of the scintillator strip (fiber is mirrored at the other end)

Scintillators were extruded at FNAL–NICADD (October–November 2012) Die for 1.9x1.5 cm 2 extrusion with bore January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)9

Fiber-and-MPPC holders and MPPCs Hamamatsu S C photosensor’s active surface January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)10 Made at ITEP

Preamplifier (one per MPPC) and carrier card designed by Hawaii (WBS 1.03) Inside detector module (7 carriers per module) Preamplifier (gain ≈ 10) Carrier (holds 15 preamps) January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)11

Radiation hardness of photosensors: OK  10 years 5 years 0.1x physics bkgd rate Light collection efficiency is unaffected by irradiation (dose equivalent to 10 years of Belle-II operation); the apparent efficiency reduction will be corrected by robust baseline measurement. MPPC intrinsic noise increases with irradiation but remains far below intrinsic background rate. noise 10 pe signal noise after irradiation January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)12

No effect was observed with proton-radiation doses 5 times higher than expected at Belle II. Radiation hardness of preamps: OK Photo-electron peaks obtained with irradiated amplifier Amplifiers gain and noise measured before (blue) and after (red) irradiation January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)13

Layer-1 module detail (looking toward front panel from interior) MPPC (hidden) to preamp Carrier to front panel (2 ribbon cables) Carrier card (1 of 7) and preamplifiers January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)14

Post-assembly-test module ready to rotate to vertical and lift (via crane) into shipping crate January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)15

Shipping crate during filling Immobilization frame to outside world January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)16

BKLM Module-Shipping Timeline in 2013: 36 modules in 4 crates (9 per crate) June JulyAugustSeptember leave VTarrive Yokohama arrive KEK Crate #1: Crate #2: Crate #3: Crate #4: VT → (truck) → Seattle → (ship) → Yokohama → (truck) → KEK using Nippon Express (“Nittsu”) as the shipping agent January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)17

Remove RPCs & Install Scintillator Modules January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)18

January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)19 Remove RPCs & Install Scintillator Modules

Post-shipment cosmic-ray measurements at KEK (September–October) Use protoboard from VT to deliver HV and power to the preamplifier carrier cards.Use two ganged power supplies (sum ~ 71V). Bring the negative preamp signal through 0.01µF capacitor to TPS 2024 oscilloscope and to LeCroy 623B discriminator. One channel at a time. Trigger = coincidence of two external scintillators. Record (geometric) efficiency, pulse height, singles rate and operating V in spreadsheet for each channel. Switch from two measurements per strip ends) to far-end-only after first 9 modules January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)20

Scintillator-assembly test stand (using cosmic rays) January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)21

Key Performance Parameters January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)22 Deliver at least 32 detector modules that each meet or exceed the threshold criteria of – at least 80% efficiency for cosmic rays, averaged over all strips in module (actual: 94%) – 15+ MPPC pixels fired for cosmic rays passing through far (mirrored) end of each scint strip (actual: ~30+ pixels) – timing resolution of at least 10 ns (actual: ~6 ns) – geometric coverage of active area: 92% in z-view, 95% in phi-view, 92% in overlap (actual: 95–96%) We delivered 36 qualified modules

Pre- and post-installation re-validations at KEK (October–November) – Off-project Use protoboard from VT to deliver HV and power to the preamplifier carrier cards. Use two ganged power supplies (sum ~ 71V). Bring the negative preamp signal through 0.01µF capacitor to TPS 2024 oscilloscope and to LeCroy 623B discriminator. One channel at a time. Count self-triggered signals less than –70mV (typical pulse height is –200mV: same as at VT). Record singles rate, pulse height, tuned operating voltage in spreadsheet for each channel. Do before and after installation of each module January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)23

Deliver electronics & firmware from Hawaii (WBS 1.03) to KEK in January Attach 2 motherboards to Layer 0 and Layer 1 modules in one of the 8 octants (either forward or backward end) Measure cosmic-ray efficiency in Layer 0 by triggering with Layer 1 – and vice versa Repeat 16 times (8 octants; forward & backward ends) Post-Installation Cosmic Ray Test at KEK (February 2014) – Off-project January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)24

Conclusion for WBS 1.04 CD-3a approved in November 2012 Delivery to KEK and post-delivery testing of installation-ready modules completed in October All modules exceeded KPPs. Post-project installation of 32 modules completed in mid-November 2013; to be commissioned in February 2014 KLM Trigger initial implementation completed in August 2013 WBS 1.04 has been completed January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)25

Backup: KLM Level-1 Trigger January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)26

Cross Sections and Level 1 Trigger rate at the design luminosity of L = 8 x cm –2 s –1 IEEE Trans Nucl Sci 58, 1807 (2011) January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)27

Level 1 Trigger inputs from detector Level 1 Trigger signal KLM Trigger January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)28

Data concentrator card (1 of 32) (WBS 1.03) Inputs from 13 RPC-readout cards in same VME crate Inputs from 2–7 scint-readout cards to KLM trigger January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)29

Universal Trigger Board (1 of ~20) is used for KLM Trigger ✓ FPGA = Virtex6 HXT ✓ Input/Output: Clock: 1 in, 3 out NIM: 10 in, 10 out 24 GTH (11 Gbps x 24) 40 GTX (6.25 Gbps x 40) LVDS: 32x2 in/out RJ45: 4 (for Belle2Link) January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)30

Leo Piilonen, US Belle II Directors Review (Pre CD-2/3) from backward Barrel FEE from forward Barrel FEE BKLM-F-S0- L* 1313 BKLM-F-S0- L1 BKLM-F-S0- L0 UT3 trigger board (32 GTX inputs) BKLM-F-S0- L* 1313 BKLM-F-S0- L1 BKLM-F-S0- L0 to Global Decision Logic EKLM-F-S0- L* 7 from forward Endcap FEE EKLM-F-S0- L* 7 from backward Endcap FEE KLM trigger finds muon tracks and K L clusters concentrators January

Algorithm finds 2D track(s) in each projection Decoder Deserialize hits Send hits in coin- cidence window to Finders x–z Finder Assemble list(s) of x -view hits in each quadrant Geometry table x–z Fitter Fit ( z, x ) hits in each list to get x 0 at I.P. straight-line fit Decision Fire trigger if ( x 0, y 0 ) is near I.P. range-dependent criterion y–z Finder Assemble list(s) of y -view hits in each quadrant Geometry table y–z Fitter UT 3 For each endcap: Fit ( z, y ) hits in each list to get y 0 at I.P. straight-line fit January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)32

Algorithm finds 2D track(s) in each projection Decoder Deserialize hits Send hits in coin- cidence window to Finders r– Finder Assemble list(s) of -view hits in each quadrant Geometry table r– Fitter Fit ( r, ) hits in each list to get r 0 at I.P. straight-line fit Decision Fire trigger if ( r 0, z 0 ) is near I.P. range-dependent criterion r–z Finder Assemble list(s) of z -view hits in each quadrant Geometry table r–z Fitter UT 3 For barrel: Fit ( r, z ) hits in each list to get z 0 at I.P. straight-line fit January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)33

Universal Trigger Module KLM Trigger hardware setup with Xilinx software environment January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)34

KLM Trigger algorithm: simulation environment on PC January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)35

Minimum period: 2.476ns (Maximum Frequency: MHz) Minimum input arrival time before clock: ns Maximum output required time after clock: 0.777ns KLM Trigger algorithm synthesis: resource utilization January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)36

Input data are available Deltas are available L0s are available 43 clock cycles 38 clock cycles Simulation of track-finding algorithm ~0.3 µs January 2014Leo Piilonen, US Belle II Directors Review (Pre CD-2/3)37