Davide Braga on behalf of the CMS Collaboration

Size: px

Start display at page:

Download "Davide Braga on behalf of the CMS Collaboration"

Donna Douglas
6 years ago
Views:

Davide Braga on behalf of the CMS Collaboration STFC

1 Beam test performance of the 2S prototype module for the High Luminosity Upgrade of the CMS Strip Tracker Davide Braga on behalf of the CMS Collaboration STFC Rutherford Appleton Laboratory Imperial College London

2 Outline Tracker upgrade & detector module The CMS Binary Chip 2 (CBC2) Mini-2S module design & testing DESY test beam Setup Results Summary & Conclusion WIT2014 2

2 Baseline design: Barrel+5Endcaps Based on 2 module types only Provides at the same

3 Phase-II upgrade of the CMS Strip Tracker 2S strip-strip double-layers ~8400 modules ~34M channels ~155m 2 PS strip-strixel double-layers ~7000 modules ~230M channels ~62m 2 Baseline design: Barrel+5Endcaps Based on 2 module types only Provides at the same time: readout data upon receipt of L1 trigger trigger Davide Braga WIT2014 3

4 Basic trigger module concept Pass Stub Fail ~2-4 mm outer sensor R inner sensor ~200μm f ~100μm High-PT tracks (stubs) can be identified if cluster centre in top layer lies within a search window in R-Φ (rows) p T cut given by: module radius (z), sensor separation and correlation window Davide Braga WIT2014 4

5 CBC2 and stub finding logic Davide Braga WIT2014 5

6 11 mm 7 mm CMS Binary Chip (CBC) 2 versions have now been produced - both in 130nm CMOS CBC1 (2011) 128 wire-bond pads, 50 mm pitch front end designed for short strips, up to 5 cm DC coupled, up to 1mA leakage tolerant, both sensor polarities binary unsparsified readout pipeline length 6.4 msec chip worked well in lab and test beam no triggering features 4 mm 5 mm CBC1 CBC2 CBC2 (January, 2013) 254 channels ~same front end, pipeline, readout approach as CBC1 bump-bond layout includes triggering features Davide Braga WIT2014 6

7 chan. mask cluster width discrimination offset correction & correlation stub shift register nearest neighbour signals CBC2 architecture 1 FE amp comp pipeline shift reg MHz diff.clock v th fast T1 trigger control fast reset test pulse v th I2C refresh test pulse v th v th bias gen OR_254 OR_stubs Ck nearest neighbour signals 256 deep pipeline + 32 deep buffer pipe. control slow control signals in blue are single-ended trig d data out stub shift reg. O/P trigger O/P I2C reset 254 channels: 127 from each sensor layer channel mask: block noisy channels from trigger logic CWD logic: exclude wide clusters >3 correlation logic: for each cluster in lower layer look for cluster in upper layer window trigger output: 1 bit per BX indicates correlation logic found one (or more) stubs triggered data out: unsparsified binary data frame in response to L1 trigger Davide Braga WIT2014 7

8 Stub finding logic Cluster width discrimination (CWD) logic exclude clusters with hits in >3 neighbouring channels wide clusters not consistent with high pt track top layer clusters zero offset offset window width Offset correction & correlation logic for a cluster in bottom layer, look for correlating cluster occurring in window in top layer window width controls pt cut stub found if cluster in bottom layer corresponds to cluster within window in top layer programmable up to ± 8 channels bottom layer clusters offset defines lateral displacement of window across chip programmable up to ± 3 channels n+1 CWD logic r f channel comparator outputs n 1/2/3 strip cluster on channel n beam n-1 programme cluster width to accept Davide Braga 8

9 S-curves number of events counts noise [e RMS ] CBC2 performance M. Raymond, IC noise & power vs. external capacitance All core functionality meets requirements Correlation functionality verified with test pulses, cosmics (backup), and in test beam Analogue performance close to simulation and specifications e.g. 1000e noise for 5 cm strips (~8 pf) achievable for total channel power of 350 uw electrons mode / measured / simulation external capacitance [pf] power/chan. [uw] fc fc comparator threshold VCTH [mv] channel offsets tuning holes mode σ=4.23 (10mV) before tune after σ=0.22 (0.5mV) ~70e comparator threshold [I2C units] Davide Braga WIT2014 9

10 CBC2 Testing Activities Davide Braga WIT

11 CBC2 testing activities Wire-bond CBC2 Useful to develop wafer probe procedures X-rays TID testing 2xCBC2 hybrid Hybrid characterization and chip integration Bump-bonded ASICs Inter-chip links & logic 2xCBC2 mini-module + sensor Sr-90 source Cosmic rays Beam Test Davide Braga WIT

8 test groups 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 12 11 10 9 8 7 6 5 4 3 2 1 11 10 9 8 7 6 5 4 3 2 1 channel channel on chip on layer 2 channel on

12 8 test groups channel channel on chip on layer 2 channel on layer 1 Results with test pulse Test pulse together with individually-programmable channel masks can be used to fully exercise the coincidence logic Davide Braga WIT

13 Logic tests using beta source Sr90 scintillator

Total Ionizing Dose test Initial Xray irradiation to

significant change in performance, moderate increase

14 Total Ionizing Dose test Initial Xray irradiation to 10 Mrads and 1wk C CBC2 operated continuously during irradiation and annealing monitored currents, biases, pedestal, noise no significant change in performance, moderate increase in current before annealing Bias/readout CBC2 Xray tube Davide Braga WIT

15 Beam Test Davide Braga WIT

Pt module beam test at DESY December 2013 4 GeV positron beam

different strip sensors Custom control and DAQ Module #3

8 mm Strip length = 50 mm #channels = 256 Module #4 (DUT) CNM,

8 mm Strip length = 54 mm #channels = 254 positron beam low

scintillator Datura telescope planes pt module ~4 GeV e+

16 Pt module beam test at DESY December GeV positron beam Datura telescope + 2 pt modules (1 fixed, 1 rotatable) with 2 different strip sensors Custom control and DAQ Module #3 (FIXED) Infineon, n-type Sensor 80x300 µm dl = 2.8 mm Strip length = 50 mm #channels = 256 Module #4 (DUT) CNM, p-type Sensor 90x270 µm dl = 2.8 mm Strip length = 54 mm #channels = 254 positron beam low angular divergence (small angular error) scintillator scintillator Datura telescope planes pt module ~4 GeV e+ module #3 on fixed assembly TOP VIEW: strip direction into page module #4 on rotation stage Mimosa26 MAPS (DATURA telescope) WIT2014

17 Pt modules & sensor variants A. Honma - CERN UPPER SENSOR 3 PT modules taken to DESY: 2 different sensor types 2.75 mm sep n one module left as backup LOWER SENSOR each CBC2 chip takes 127 inputs from upper sensor and 127 inputs from bottom sensor

Modules taken to DESY Each module assembled into identical units for mechanical

interface card - light-tight aluminium cast ~10x20cm boxes for support and

20 via external Arduino-based controller - connectors for DAQ, power, temperature

18 Modules taken to DESY Each module assembled into identical units for mechanical support, environment control & handling Peltier hybrid temperature sensor 2CBC2 interface card - light-tight aluminium cast ~10x20cm boxes for support and heatsinking - aluminium foil windows - Peltier elements to actively cool hybrid to 20 via external Arduino-based controller - connectors for DAQ, power, temperature control, HV bias LV power DAQ bias, temperature control M. Raymond, IC Davide Braga WIT

19 Beam test DAQ M. Pesaresi, IC L. Gross, IPHC DAQ PC RJ-45 LVDS mezzanine scintillator trigger Gb Ethernet GLIB external trigger backpressure DESY TLU Bristol x1 FMC GLIB SFP GLIB GLIB SFP GBT optical fibres (max 4/8) to DATURA telescope <5m control and DAQ based on CERN GLIB emulating GBT functionality prototype modules Davide Braga WIT

20 Beam test DAQ (2) BE GLIB GbE - basic elements for the CMS based DAQ were put in place just in time for beam test GLIB Streamer GLIB Supervisor RU BU data unpacker DIGIS + RAW DIGIS online analysis package FU SM DQM ~1.2cm strips WIT

Thresholds scan at normal incidence Trigger and stub latency scan Beam test issues - most issues solved during week s running: DAQ ran stably for last few days (albeit at ~300Hz) - some issues that

data sets based on hit patterns Commissioning steps much time was dedicated to commissioning the modules in beam Measurements 24h shifts in the last 3-4 days over 120M events to disk!

21 Thresholds scan at normal incidence Trigger and stub latency scan Beam test issues - most issues solved during week s running: DAQ ran stably for last few days (albeit at ~300Hz) - some issues that took a while to solve to do with the external TLU Timestamp issue with the Datura Telescope: BX and Orbit number not saved by the GLibStreamer Ali Harb (DESY) looking into ways to synchronise the two data sets based on hit patterns Commissioning steps much time was dedicated to commissioning the modules in beam Measurements 24h shifts in the last 3-4 days over 120M events to disk! threshold scans at normal incidence (high statistics, fine VCTH steps) low threshold scan (to see noise floor) latency scans angular scans with high statistics, nominal thresholds threshold vs angular scans (low statistics, coarse VCTH steps) runs with different cluster width discrimination settings dedicated runs to check DAQ and telescope synchronisation module #4 bottom sensor

22 P T cut principle demonstrated M. Pesaresi, IC S. Viret, IPHC EXP. p T cut: 2.14 GeV/c FIT: p T cut: 2.17 GeV/c σ: 0.1 GeV/c Pt Selection cut: simulation measured efficiency CBC2 correlation window was set to +/- 7 strips; 0 strip offset reconstructed p T cut of r=75cm layer P T -cut reconstructed from beam test data matches the design one exactly Sharp turn-on First experimental result to prove the stub selection concept! Davide Braga WIT

23 Stub Efficiency vs. Angle Reconstructed stub Efficiency vs. P T r = 0.6m r = 1.1m Kirill Skovpen, IPHC Davide Braga WIT2014

24 Fraction of total stubs Average cluster width [strips] Fraction of total stubs '1strip' '2strip' '3strip' '4strip' '5strip' '6strip' '7strip' '8strip' '9strip' '10strip' Angle of incidence [degrees] Angle of incidence [degrees] Study of cluster width 10 0 Average width and cluster width distribution study '1strip' '2strip' '3strip' '4strip' '5strip' '6strip' '7strip' '8strip' '9strip' '10strip' Data used for comparison with CMSSW Digitizer reconstruction (Suchandra Dutta and Suvankar Roy Chowdhury, Saha Institute of Nuclear Physics) Angle of incidence [degrees] 25 30

25 Strip noise Martin Delcourt, UC Louvain Noise probabilities per CBC and per sensor Noise obtained by fitting a binomial to the number of hits per strip with no beam One case (CBC1-sens.0) not well described by single binomial need more data Noise smaller than 10-6 per strip per trigger Davide Braga WIT

26 Martin Delcourt, UC Louvain Detection Efficiency ε = 99.91±0.01% (CBC0) 99.89±0.01%(CBC1) Loss of efficiency on the sides Loss of efficiency in the due to bad strip

27 Module alignment Kirill Skovpen, IPHC Martin Delcourt UC Louvain φ = 2.1 ± 0.6 φ = 2.17 ± 0.03 Davide Braga WIT

28 Clean Event Scan Martin Delcourt UC Louvain VCTH ~ 100 optimum setting for low noise and clean events Proportion of events with cluster occupancy = 1 for all sensors. If occ=1 for all sensors, proportion on events with distance > 4 between hits on sensor 0 and sensor 1.

Module-Test Setup at CERN Fully equipped Module-Test Setup available at CERN

utca shelf infrastructure for testing, development & debugging mechanics for

integration tests of components, development of commissioning & calibration

29 Module-Test Setup at CERN Fully equipped Module-Test Setup available at CERN exactly the same DAQ chain as in the beam test (FW & SW) fully integrated in utca shelf infrastructure for testing, development & debugging mechanics for testing with cosmics / radioactive source & scintillator trigger used for: integration tests of components, development of commissioning & calibration procedures currently focusing on: CBC hit correlation studies Georg Auzinger, Stefano Mersi, CERN

30 Conclusions and Future Work Davide Braga WIT

Conclusions and Future Work Two successful full-size prototypes of new Outer Tracker ASIC and mini-2s module in hand CBC2 working to specs, stub finding logic functioning First beam test of the 2S

31 Conclusions and Future Work Two successful full-size prototypes of new Outer Tracker ASIC and mini-2s module in hand CBC2 working to specs, stub finding logic functioning First beam test of the 2S prototype module successful commissioning & readout of two prototype stacked modules in beam, prompt analysis and online s/w data taken for measurement of sensor/front-end performance & to demonstrate ability of 2S modules to discriminate on pt DAQ and readout system integrated, commissioned & working in beam test TID testing started, no problem so far Goals for 2014: Continue Pt Module characterization in beam (sensor + front-end ASIC + hybrid + DAQ) Applied for beam time at the end of the year at SPS First characterization of irradiated sensors (cooling required) 8-CBC2 hybrids for the telescope planes Validation of full DAQ chain, including future ASIC (Concentrator) emulated in FPGA First results to prove the track-trigger concept CMS ambitious plans for a track trigger look promising A. Honma, CERN

Acknowledgements CBC2: RAL: Davide Braga, Lawrence Jones, Peter Murray, Mark Prydderch IMPERIAL COLLEGE: Geoff Hall, Mark Pesaresi, Mark Raymond CERN: Federico Faccio, Kostas Kloukinas, Stefano

32 Acknowledgements CBC2: RAL: Davide Braga, Lawrence Jones, Peter Murray, Mark Prydderch IMPERIAL COLLEGE: Geoff Hall, Mark Pesaresi, Mark Raymond CERN: Federico Faccio, Kostas Kloukinas, Stefano Michelis 2S module: CERN: Georges Blanchot, Alan Honma, Mark Kovacs, Francois Vasey Beam test and analysis: CERN: Georg Auzinger, Stefano Mersi DESY: Ali Harb, Andreas Mussgiller, Doris Eckstein HEPHY: Thomas Bergauer, Wolfgang Treberspurg IPHC: Christian Bonnin, Kirill Skovpen, Laurent Charles, Laurent Gross KIT: Alexander Dierlamm, Martin Printz IMPERIAL COLLEGE: Davide Braga, Jonathan Fulcher, Mark Pesaresi, Mark Raymond SAHA INSTITUTE OF NUCLEAR PHYSICS: Suchandra Dutta, Suvankar Roy Chowdhury UC LOUVAIN: Christophe Delaere, Jerome De Favereau, Martin Delcourt UNIVERSITY OF BRISTOL: David Cussans, Fionn Ball

The CMS Binary Chip for microstrip tracker readout at the SLHC

The CMS Binary Chip for microstrip tracker readout at the SLHC OUTLINE brief review of LHC strip readout architecture CBC design and measured performance first test beam results future directions summary