1. 06 February 2007Fabrizio Cei1 INFN and University of Pisa INFN Scientific Commission I Rome, 06 February 2007 Status of MEG Software

2. 06 February 2007Fabrizio Cei2 Outline  MEG Software organization  Status of Monte Carlo simulation  Short remind of analysis framework  Status of analysis codes  Status of MEG computing power @PSI  Preliminary analysis of 2006 beam test

3. 06 February 2007Fabrizio Cei3 MEG Software Organization MC Bartender(ROME) Analyzer(ROME) DAQ ZEBRA ROOT MIDAS ROOT Simulation Real Data WFM andsimulation WFM and pile-up simulation

4. 06 February 2007Fabrizio Cei4 Status of Monte Carlo MEGMC program - written in Geant3.21; - data output in ZEBRA banks, automatically converted to C++ structures (readable from analysis codes); to C++ structures (readable from analysis codes); - it simulates pair (e.g.  e  ) or single (e.g. Michel positrons) events; - full simulation of the detector: DCH, TICP, TICZ, LXe, Target, Beam, Magnet and magnetic field, mechanical supports, Beam Beam, Magnet and magnetic field, mechanical supports, Beam Upstream and Downstream elements, calibration devices (NaI, Upstream and Downstream elements, calibration devices (NaI, proton accelerator tube & targets, LH 2 target,  sources …); proton accelerator tube & targets, LH 2 target,  sources …); - LXe & TC beam tests simulated by specific modules (tbeam & tbtc); - LXe & TC beam tests simulated by specific modules (tbeam & tbtc); - different running configurations (final or run 2006); - large samples (~ 10 5 ) of events generated without difficulties; - few event types missing, e.g. LED Completion Status : 99 %

5. 06 February 2007Fabrizio Cei5 Examples of MC events  e  event Michel positron RUN 2006 configuration: No LXe, no TICZ, 8 DCH, TICP displaced by ± 12 cm

6. 06 February 2007Fabrizio Cei6 MC Man Power All persons @ 10  50 % of their time All persons @ 10  50 % of their time:  Coordination: S. Yamada (UCI), F. Cei (Pisa);  SVN repository: S. Yamada, P. Cattaneo (Pavia);  Event generation:F. Cei, S. Yamada, Y. Hisamatsu (Tokyo);  LXe: S. Yamada, F. Cei, G. Signorelli (Pisa);  TICP/TICZ: P. Cattaneo, Y. Uchiyama (Tokyo);  DCH: H. Nishiguchi (Tokyo), M. Hillebrandt (PSI);  Beam & Magnet: W. Ootani (Tokyo);  Target: V. Tumakov (UCI);  NaI: Y. Nishimura (Tokyo);  Calibrations: F. Cei

7. 06 February 2007Fabrizio Cei7 Software framework: ROME ROME (Root based Object Oriented Midas Environment) is a framework generator; framework generator; ROME is separated in one detector independent and one detector ROME is separated in one detector independent and one detector dependent part. It has been adopted by other experiments; dependent part. It has been adopted by other experiments; It uses only 6 different C++ objects; It uses only 6 different C++ objects; ROME makes the “dirty job”: ROME makes the “dirty job”: creating the structure, defining C++ classes, writing many include creating the structure, defining C++ classes, writing many include files, creating the dependences and the hierarchy …; files, creating the dependences and the hierarchy …; the users and detector experts perform the “smart job”: the users and detector experts perform the “smart job”: writing the analysis methods (tasks) and the related folders writing the analysis methods (tasks) and the related folders (data stored on memory) and trees (stored on disks); (data stored on memory) and trees (stored on disks); the most important feature is the modularity: the most important feature is the modularity: the tasks can be exchanged at runtime. the tasks can be exchanged at runtime. Main developer: M. Schneebeli (PSI)

8. 06 February 2007Fabrizio Cei8 ROME Interconnections Folders Tasks Fill Read Trees Fill Flag Histograms Fill Disk (Output) Write (ROOT) Disk (Input) Read (Format: ZEBRA, MIDAS & ROOT)

9. 06 February 2007Fabrizio Cei9 ROME Event Display (ARGUS) Display includes tracks and energy deposits

10. 06 February 2007Fabrizio Cei10 Waveform/track display Liquid Xenon Drift Chamber Both used in on-line and analysis too

11. 06 February 2007Fabrizio Cei11 Status of analysis codes 1) MEGBartender  Event mixing tested; multiple formats allowed (MIDAS, ROOT, ZEBRA); allowed (MIDAS, ROOT, ZEBRA);  Waveform simulation completed for LXe, TICP/TICZ, DCH wires; work needed TICP/TICZ, DCH wires; work needed for pad simulation; for pad simulation;  Preliminary trigger simulation included; Completion status: not far from end

12. 06 February 2007Fabrizio Cei12 Status of analysis codes 2) MEGAnalyzer LXe (R. Sawada (Tokyo), G. Signorelli, Y. Uchiyama,  LXe : (R. Sawada (Tokyo), G. Signorelli, Y. Uchiyama, S. Yamada, F. Cei) S. Yamada, F. Cei) - Waveform decoding implemented; -Charge-based reconstruction algorithms implemented; most of them tested; -Timing reconstruction & calibration algorithms under implementation and testing; under implementation and testing; - Peak finding and pattern recognition tasks existing; Completion status: ~ 90 %

13. 06 February 2007Fabrizio Cei13 Status of analysis codes 3)  TICP/Z (P. Cattaneo, Y. Uchiyama, D. Zanello (Rome) F. Xiao (UCI), A. Barchiesi (Rome), S. Dussoni (Genova)) F. Xiao (UCI), A. Barchiesi (Rome), S. Dussoni (Genova)) - Waveform analysis implemented (baseline, pulses, charges, left/right timing, pile-up flags …); charges, left/right timing, pile-up flags …); - Preliminary hit reconstruction implemented (Q,t L,t R )  (z, ); (Q,t L,t R )  (z, ); - TBD: correlation between adjacent bars, correlation with DCH correlation with DCH Completion status: ~ 75 %, but sufficient man power

14. 06 February 2007Fabrizio Cei14 Status of analysis codes 4)  DCH (H. Nishiguchi, M. Schneebeli, M. Hildebrandt) - 3D-map of magnetic field; - 3D-map of magnetic field; - tracking by (preliminary) Kalman filter implemented; - tracking by (preliminary) Kalman filter implemented; - waveform decoding existing, but slow; - waveform decoding existing, but slow; - hit extraction (x,t  3D coordinates) still missing; - hit extraction (x,t  3D coordinates) still missing; - extraction of z-coordinate from cathode pad - extraction of z-coordinate from cathode pad information still missing. information still missing. Since the DCH software seems the most critical part, (some tasks missing, reduced man power) we decided to put some effort on this.

15. 06 February 2007Fabrizio Cei15 Status of analysis codes 5)  Trigger (G. Signorelli, D. Nicolò (Pisa)) - Trigger info/banks coded in MIDAS: run #, event #, trigger code, - Trigger info/banks coded in MIDAS: run #, event #, trigger code, trigger WFMs, scalers (useful for determining run/live time); trigger WFMs, scalers (useful for determining run/live time); - Charge and timing reconstruction algorithms implemented - Charge and timing reconstruction algorithms implemented and under testing (see later) and under testing (see later)  Database (R. Sawada) - Two databases: MySQL and sqlite3; easy conversion. - Two databases: MySQL and sqlite3; easy conversion. - MySQL needs network; sqlite3 for stand-alone environment in a - MySQL needs network; sqlite3 for stand-alone environment in a separate svn module (megdb). separate svn module (megdb). - Included: geometry, trigger/hardware configuration, run table, - Included: geometry, trigger/hardware configuration, run table, physical constants, reconstruction coefficients … physical constants, reconstruction coefficients … Completion status: ~ OK

16. 06 February 2007Fabrizio Cei16 MEG computing @PSI Sun Fire x4100 quad core 4 GB Fiber Channel Switch 15 x 500 GB SATA GBit Ethernet Offline cluster for MEG Presently available : 20 CPU cores + 30 TB disk Final situation : 64 CPU cores + 100 TB disk in totalPresently available : 20 CPU cores + 30 TB disk Final situation : 64 CPU cores + 100 TB disk in total Easily extensibleEasily extensible GBit link to online cluster requestedGBit link to online cluster requested

17. 06 February 2007Fabrizio Cei17 Preliminary analysis of 2006 run Hardware configuration: - no LXe; no TICZ; 8/16 DCH installed; - TICP shifted by ± 12 cm for mechanical compatibility; - Both detectors equipped with trigger and read-out electronics (Domino sampling chip); and read-out electronics (Domino sampling chip); - trigger by single TICP Bar  DCH events seen as “accidentals” in time as “accidentals” in time MC Simulation: - 500000 Michel positrons with isotropic angular distribution and “RUN2006” configuration distribution and “RUN2006” configuration

18. 06 February 2007Fabrizio Cei18 Run 2006 configuration Only 3 DCH effectively working; one broken wire in one of these. No couples of adjacent chambers working  no tracks. Six TICP bars always off (four of them equipped with test electronics); two others off in some runs DS US

19. 06 February 2007Fabrizio Cei19 Pulse finding algorithm  Designed to be fast (no fit, ~ 8 kWFM/s) and versatile; and versatile;  Smooth the waveform to reduce noise;  Find reasonable baseline.

20. 06 February 2007Fabrizio Cei20 Search for pulse  Bool_t inPulse + FindPulseStart() and FindPulseEnd();  Pulse Merge / Pulse Cleanup (remove too small pulses).

21. 06 February 2007Fabrizio Cei21 Application to DCH 1) DCH waveform example Wire signals Cathode signals Cathodes not yet used Low Rate High Rate Time correlation with the trigger signal DRS Data (500 MHz)

22. 06 February 2007Fabrizio Cei22 Application to DCH 2) Plane A Plane B  Events selected by wire-end coincidences coincidences  Cuts on MC events by comparing expected/measured comparing expected/measured charge distributions charge distributions (~ 30 % uncertainty) (~ 30 % uncertainty)  Reasonable agreement with MC predictions MC predictions Slightly lower rate on Plane B (outmost) A B COBRAcenter Black Data Red MC

23. 06 February 2007Fabrizio Cei23 Application to DCH: 2D-map In the MC we have the hit z-coordinate, for data we use asymmetries In the MC we have the hit z-coordinate, for data we use asymmetries A “shadow” is present on the upstream side A “shadow” is present on the upstream side

24. 06 February 2007Fabrizio Cei24 Application to TICP 1) Amplitude vs Charge relation Trigger threshold: 50 mV in amplitude for both channels && 200 mV on sum. Amplitude vs charge relation fitted for determining an equivalent charge threshold and simulate the trigger threshold for MC events. DRS amplitude attenuated by a factor five

25. 06 February 2007Fabrizio Cei25 Application to TICP 2) Charge distributions Bar 6, run 154 Red Data Black MC (scaled to superimpose peaks) Bar 11, run 154 MC events selected by cuts on charge distribution to schematize trigger selections (25% uncertainty) All bars together. Individual calibrations to be improved

26. 06 February 2007Fabrizio Cei26 Application to TICP 3) Run 154 Low Intensity Red Data Black MC Run 236 High Intensity Number of hit bar per event

27. 06 February 2007Fabrizio Cei27 Number of hits on each bar Down stream Up stream Run 154 Low Intensity Run 236 High Intensity Red Data Black MC

28. 06 February 2007Fabrizio Cei28 DCH and TICP rates Expected rates computed by using the nominal opening of beam slits. Measured rates computed by using MC evaluated efficiencies:  DCH (single dch)  6.6 %  DCH (single dch)  6.6 %  TICP  4.2 %  TICP  4.2 % Estimated uncertainties on trigger rates  20  30 %: - limited statistics; - limited statistics; - MC cuts - MC cuts  Rate (x 10 7  /s) 16% 65% 88% 100%

29. 06 February 2007Fabrizio Cei29 Trigger data 1) WFM TICP

30. 06 February 2007Fabrizio Cei30 Trigger Data 2) Correlation with DRS data Correlation with DRS data DRSCharge TriggerCharge

31. 06 February 2007Fabrizio Cei31 Trigger Data 3) 10 mV threshold imposed by analysis program (otherwise, the pulse finding algorithm becomes meaningless). Charge distribution for data go down to very small values, as predicted by MC simulation. Low Charge distribution (accidental events collected in 5  s time window Low Charge distribution (accidental events collected in 5  s time window ) Red MC Black Data

32. 06 February 2007Fabrizio Cei32 Results of TICP tests 1)

33. 06 February 2007Fabrizio Cei33 Results of TICP test 2)

34. 06 February 2007Fabrizio Cei34 TICP test bar results

35. 06 February 2007Fabrizio Cei35 Conclusions  The MEG software is in an advanced state of preparation: preparation: –MEGMC + MEGBartender: almost finished –MEGAnalyzer: LXe: close to completion; TICP/Z: not very far from end; sufficient effort; TICP/Z: not very far from end; sufficient effort; DCH: delayed; more effort needed Trigger/Database: ~ ok. Trigger/Database: ~ ok.  The analysis of RUN 2006 beam test was entirely performed within the general framework, which performed within the general framework, which looks adequate for the whole MEG analysis. looks adequate for the whole MEG analysis.

36. 06 February 2007Fabrizio Cei36 Trigger data 2) Timing RUN 154 (slit at 16 %) RUN 236 (slit at 100 %) Time correlated with trigger pulse 5 microseconds window

37. 06 February 2007Fabrizio Cei37 Trigger data 2) Charge (T 1 -T 2) vs Log(Q 1 /Q 2 ) 10 ns time binning (poor resolution); however, well visiblecorrelation well visible correlation Time and charge obtained through fitting operation:

38. 06 February 2007Fabrizio Cei38 Application to TICP 4) Log(Inner/Outer Charge) Distribution (  z)  T vs Log(In/Out Charge) Red Data Black MC Rough timing algorithm (it selects the pulse start in one 2 ns bin), but reasonable correlation