Скачать презентацию MEG 2008 Run x E Ma E e Скачать презентацию MEG 2008 Run x E Ma E e

ba6cb5429df054e8e1ee7e3e3fa516a2.ppt

  • Количество слайдов: 19

MEG 2008 Run x E Ma E e/ Run Coordinator’s view Peter-Raymond Kettle MEG MEG 2008 Run x E Ma E e/ Run Coordinator’s view Peter-Raymond Kettle MEG Review February 2009 1

Situation - Review Feb. 2008 Back to “Square One” - TOTAL Detector DISMANTLED post Situation - Review Feb. 2008 Back to “Square One” - TOTAL Detector DISMANTLED post 2007 Engineering Run - for Maintenance/Repair /Improvement TCs: Fibre light-leak + new N 2 Bags + APD amplifier/electronics TCs DCs: Support Structure + new Target Angle + HV Investigation EC BTS Peter-Raymond Kettle Calo: new HV Feed-thro’s + LN 2 Cooling-pipe mod. (heat-load) + lnvest. LXe Light-Yield mod. Purification system C-W DS-EC + IS Calo. MEG Review February 2009 2

2008 Beam Time & Constraints 2008 Run Goals: final beam operational conditions to be 2008 Beam Time & Constraints 2008 Run Goals: final beam operational conditions to be tuned/optimized {( -, - & p-beams C-W)} Full set of detector calibrations + optimization/exploitation of various techniques optimized detector/trigger settings as fuller set of information as possible, necessary for Data-analysis + “Long-term” Goal of understanding our Detector End-of-Shutdown Max. Expectation MEG-Physics Data ~ 16 weeks Peter-Raymond Kettle MEG Review February 2009 3

Planning & Organization Basic Run Layout Parasitic Run Beam optimization in parallel with Debug, Planning & Organization Basic Run Layout Parasitic Run Beam optimization in parallel with Debug, Tune & Calibrate Full Run Part I CEX Run + Trigger Setup + calibrations Detector monitoring Full Run Part II Pre-physics data check Physics Data (MEG + RD) Run Coordinator 2 Shift Coordinators 7 weeks tot. 13 Shift Coordinators 2 weeks/person Parasitic Run Total of 62 persons for 959 shifts (Full Run Only) to allow for flexibility + continuity: Staggered & Overlapping shift system Daily Run Meetings (on-site) Weekly Video Run Meeting (Collaboration-wide) later, weekly Video Physics Analysis Group Meeting Web-based Schedule + Shift list + “On-call” List Peter-Raymond Kettle MEG Review February 2009 12 Hr Shifts: 1 DAY SHIFT (Beam Group) 10: 00 – 22: 00 1 NIGHT SHIFT max. 22: 00 – 10: 00 1 Shift Leader Nights Manned by Detector Experts Full Run 8 Hr Shifts: 1 DAY SHIFT 1 EVENING SHIFT 1 NIGHT SHIFT 07: 00 – 15: 30 15: 00 – 23: 30 23: 00 – 07: 30 1 Shift Leader + 1 Crew Member 4

Organization In Practice Detailed Information Access • • “How to” information database for shift Organization In Practice Detailed Information Access • • “How to” information database for shift crews on MEGWiki • fully searchable & cross-referenced Electronic Logbooks for all sub-detectors & Run shift-crews ts e he experiment/instrument • easy web-based S control 5 shift-crews for (e. g. Beam Line – magnets, separator) • comprehensive electronic Hardware/Software check-list for shift crews t Lis t hif S ES L P AM X E push-button changing of rate with auto magnet/slits hysteresis cycling le du e ch S Peter-Raymond Kettle MEG Review February 2009 5

Detector Synopsis TCs DCs TCs: • Fibres working, problem DAQ control of DS fibres Detector Synopsis TCs DCs TCs: • Fibres working, problem DAQ control of DS fibres • Laser temp. control problems Calo: • HV feedthroughs replaced • Liquid & Gaseous purification success LY-behaviour needs further study + PM gain stability air-doping still persist!!! Na. I: C-W: • proved “Essential Tool” Li (17. 6, 14. 6 Me. V) + B (4. 4, 11. 7, 16. 1 Me. V) lines – Energy + Timing Calo. Peter-Raymond Kettle DCs: • HV instability problems with C-W MEG Review February 2009 • New APD preamps • automated mover • temp. contolled APDs • E/E ~5 -6% ( ) Na. I 6

Trigger + DAQ Trigger + Splitters Online Cluster Megonxx Multi-trigger implementation: Final Complement of Trigger + DAQ Trigger + Splitters Online Cluster Megonxx Multi-trigger implementation: Final Complement of 29 Triggers implemented multiple & pre-scaled (MEG=11, RMD=5) Single & Coincidence detector triggers crucial for monitoring/Calibration DRS v 2 + part v 3 e. g. 7 Li, 11 B C-W, , CR, LED – RMD, 0, 0 -Dalitz CEX software collimators LXe responce Direction matching e (planned TC fibres) too slow still XEC PMT-index + TC-bar(index, z) where z from bar charge-ratio Trig. Monitoring via (cyclic-buffers) lcmeg 05 lcmeg 04 DRS 3 – partly implemented (clock signals, temp eff. Etc. ) Limits: • DAQ/DRS readout limited by VME (83 MB/s) ~ 30 events/s full waveforms (threading) • Online (backend) 2 TB storage • Offline (lcmeg) 64 CPUs + 104 TB disk • “Lazylogger” autocopy Online Offline factor 2 compression offline Peter-Raymond Kettle MEG Review February 2009 lcmeg 03 lcmeg 02 lcmeg 01 Offline Cluster lcmeg 7

Arsenal of Standard Calibration Tools 7 Better t, makes it possible to take data Arsenal of Standard Calibration Tools 7 Better t, makes it possible to take data with higher beam intensity A few days ~ 1 week to get enough statistics 0 - + p 0 + n 0 (55 Me. V, 83 Me. V) - + p + n (129 Me. V) LED Laser NO TY ET ST AN DA RD radiative decay e Lower beam intensity < 10 Is necessary to reduce pileups n n (rough) relative timing calib. < 2~3 nsec Laser PMT Gain Higher V with light att. Can be repeated frequently 10 days to scan all volume precisely (faster scan possible with less points) e+ LH 2 target alpha MEG Detector Standard Calibrations e- Proton Acc Li(p, )Be Li. F target at COBRA center 17. 6 Me. V Bi Tl F Li(p, 1) at 14. 6 Me. V ~daily calib. Li(p, 0) at 17. 6 Me. V Peter-Raymond Kettle Cold GXe LXe K PMT QE & Att. L Can be used also for initial setup Nickel Generator off on quelle Illuminate Xe from the back OT N ET 3 cm Y Source (Cf) transferred by comp air on/off MEG Review February 2009 D RMe. V Nickel γ-line 9 DA AN ST 20 cm Na. I Polyethylene 0. 25 cm Nickel plate 8

New & Improved Calibration Techniques Multiple calibration techniques proved Essential for deconvoluting complex effects New & Improved Calibration Techniques Multiple calibration techniques proved Essential for deconvoluting complex effects >16. 1 Me. V >11. 7 Me. V 4. 4 Me. V 11. 7 & 4. 4 Me. V s Coincident in time (94%) & no angular correlation Li used for E-calibration, B can be used for Δtabs(LXe-TC) or Δt(TC-TC(inter-bar)) Energy deposit in XEC (1) New Lithium Tetraborate. Target (Li 2 B 4 O 7) for C-W: - advantage both Li- & B-lines simultaneously available without large X-section of F (~ 6 -7 Me. V), from Li. F 4. 4 and 11. 6 Me. V Compton Edges “Energy” deposit in TC (2) New timing calibration technique during CEX: - use Dalitz decay for intercalibrating LXe & TC detectors by tracking e+ in DCs ( 0 → e+e- ) used successfully for measuring absolute Δt(LXe-TC) of reference TC-bar, can then intercalibrate bar using Boron e. g. LY (3) Am/Be – neutron source as a source Am/Be ≡ Li of 4. 44 Me. V Gammas from 2+ state of 12 C* via 9 Be(α, n)12 C Peter-Raymond Kettle MEG Review February 2009 9

Calibration Techniques cont. (4) Use of tuned monochromatic positron beam being investigated as a Calibration Techniques cont. (4) Use of tuned monochromatic positron beam being investigated as a means of e. g. studying our positron spectrometer tracking efficiency vs. emission angle or momentum, with high statistics, in a momentum range equivalent to real MEGconditions!!! Mechanism: positron-Nucleus elastic scattering from light nuclei at around Pe ~ 50 Me. V/c “Coherent” nuclear recoil, spin-effects, magnetic terms all ~ negligable nuclear form-factor introduces a small effect X-sections “well known” basically “Mott-scattering” Reality: • MEG beam can be tuned to give ~ 50 Me. V/c e+ with a max. rate of ~ 8· 108 e+/s at 2 m. A proton current with ΔP/P ~ 7% FWHM obviously would reduce ΔP/P to achieve “monochromaticity” though at the cost of rate. • Wien-filter does not work at this momentum to sufficiently separate e+ from + but a 2 mm CH 2 -degrader at the collimator system in front of BTS DOES! Carbon target ρ ~ 2. 1 g/cm 3 t < 1 cm thick, and 107 e+/s Integrated X-section: 30° < Δθ < 120° & Δφ = 2. 5 mbarn ~ 1300 events/s Peter-Raymond Kettle MEG Review February 2009 10

2008 Run Conditions (1) New Target Angle: - modification DC Support Structure optimal @ 2008 Run Conditions (1) New Target Angle: - modification DC Support Structure optimal @ ~ 21° to match beam stopping distribution etc. Prior 2008 limited by DC Support structure to max. 13° • Conventional • Photogrammetric (outside COBRA) • Photogrammetric (inside COBRA) = (20. 6 ± 0. 2)° - target inclination angle = (20. 4 ± 0. 2)° = (20. 3 ± 0. 3)° Target Inclination 2008 = (20. 5 ± 0. 3)° (2) Beam Intensities: - apart from “Normal” beam intensity 2 further tunes were optimized based on standard degrader 300 m Mylar – “Ultra-low” & “High” Mode R Measured Rate COBRA at 2 m. A Rstop Stopping Rate at 2 m. A (εSTOP= 0. 794) “High” 8. 4 107 +s-1 6. 7 107 +s-1 “Normal” 3. 5 107 +s-1 ~ 2. 8 107 +s-1 “Ultra-low” 1. 5 106 +s-1 ~ 1. 2 106 +s-1 Measured values at 7% air contamination 1% Air ~ 10 m Mylar degrader - Not compensated for in 2008!!! Peter-Raymond Kettle MEG Review February 2009 11

Run Conditions cont. (3) COBRA He-Concentration: - for DC HV-stability reasons air-doping of COBRA Run Conditions cont. (3) COBRA He-Concentration: - for DC HV-stability reasons air-doping of COBRA Environment was necessary Physics Run Classification: MEG Data ONLY(before/after DC COBRA test) 96% He 95% He P 61, P 70, P 63 63 O 2 -sensor 61 COBRA 60 Part 1: 93% Part 2 Part 1 P 61 blue P 70 green P 63 red P 60 light blue Mean Air-doping (Physics Run Part 1 + 2) = 6% Peter-Raymond Kettle 70 62 11/9 92% US-EC DS-EC MEG Review February 2009 (0100) – 20/10 (0100) 35 days Part 2: 27/10 (1149) – 06/11 (2359) – 23/12 11 + 32. 5 days 1% Air ~ 10 m Mylar degrader Not compensated for!!! 12

2008 Beam Time Influences 3 Major factors influenced the maximizing of the available beam 2008 Beam Time Influences 3 Major factors influenced the maximizing of the available beam time for Physics Data-taking – such that substantial extra investigation time was necessary (1) Calorimeter: 2007 Light-yield << expected both for s & s (Q/A) /(Q/A) ~ 1. 25 expect LP~ 1. 92!!! Light-yield + PMT gain driftinstalled 2008 Contamination? new purifier • Calorimeter – Significant time was invested with monitoring/understanding of LY vs. purification time • Electronics Noise + Baseline stability Using C-W Li, CR, s & LEDs • Drift Chambers - HV stability Liq. P → Induced Noise on electronics Liquid & Gaseous & No purification scenarios studied Gas. P → minimal Noise Questions to answer: • can one survive without Liq. P for 3 weeks between inter-accelerator shutdowns and only rely on Gas. P? • What happens to LY without any purifications? Initial Purification May 2008 L-Y - 0. 7%/5 days s also CR same response! s Liq. P Peter-Raymond Kettle Gas. P No. P MEG Review February 2009 13

Beam Time Influences Calo. - cont. • Variations in no Photo-electrons seen…LY changing + Beam Time Influences Calo. - cont. • Variations in no Photo-electrons seen…LY changing + ? • PMT- Gain variation seen vs. Beam rate drastic during CEX changes of several % possible!! Stable at low rate • PMT Gain well monitored using LEDs …Reason for instability? in principle effect already compensated for “zener diodes”!!! could this be aging? • Thus frequent LED calibrations used as gain normalization for light/energy measurements • time constants for rate-changes measured (beam-blocker) • therefore in principle all ingredients available for corrections PM Gain CEX - Off Norm Beam 1/2 LY nearly reached optimal value LY is constant LY > CEX 2007 Level Peter-Raymond Kettle How does this affect our energy scale Extrapolated for low-energy low-rate? MEG Review February 2009 14

Effect on Energy Scale Rate-dependent effect B-B “opened” ~30 mins. For our Photon Energy-scale Effect on Energy Scale Rate-dependent effect B-B “opened” ~30 mins. For our Photon Energy-scale we extrapolate from “Low-energy” “Low-rate” C-W data to “High-energy” “High-rate” CEX pion data What is the rate dependency at CEX-rates? Not enough LED data taken during intial CEX Hence new “mini-CEX” run at end of December B-B “closed” CEX ~ 4% discrepancy from Extrapolation to CEX energies before correction from “Mini-CEX” After? better but not perfect! More work needed Peter-Raymond Kettle MEG Review February 2009 Before Correction C-W 15

Beam Time Influences – cont. (2) Electronics: baseline instability as well higher noise content Beam Time Influences – cont. (2) Electronics: baseline instability as well higher noise content on DRS i/ps 2 reasons found – (i) i/p stages (diodes + resistors)of splitter damaged by sparking from defective Calorimeter feed- thro’s. all boards modified with new higher rated diodes & resistors changed (ii) burst-noise suppressed with external shielding of flat-band calorimeter cables between splitter & DRS external shielding added to all calo-cables between splitter & DRS (3) Drift Chambers: HV-stability of chambers persists, seems to be a time dependency before onset & seems worsened by CEX pion beam then worsens with time. Gives a complicated time-dependent e+ detector efficiency Air doping + overpressure + gas-mixture investigated during dedicated combined electronics/ Calo. /DC maintenance week 100% Anode Hit-map. He-Con. C. ΔP(DC-COBRA). 0% Peter-Raymond Kettle MEG Review February 2009 16

Beam Time/Data In view of the complex & overlapping problems that were studied & Beam Time/Data In view of the complex & overlapping problems that were studied & monitored during the “Parasitic” & Part 1 Phase of the “Full Run” the following schedule evolved necessitating a mini-CEX at the end of the period to evaluate the rate dependency during the full CEX, so that the Calorimeter Energy-scale could be fully determined: Parasitic Run: 19 th May- 3 rd July ~ 7 weeks MEG 2008 Run ( 4. 5 weeks) Beam Tests/Tuning DATA Taken Full Run Part 1: 11 th July – 31 st August ~7 weeks TBytes CEX 21 st July – 31 st August (6 weeks) Full Run Part 2: 1 st September – 23 rd December ~16 weeks Pre-Physics Data (~ 3 weeks) Physics Data Part 1 35 Days 139 TB MEG Maintenance/Repair ~ 7 Days Physics Data Part 2 43. 5 Days Mini-CEX ~ 7 Days DATA MEG (Runs# 23987 - 40997) 10859 Runs a 2 k events 22. 4 M Triggers Time 49: 18: 50: 49 RMD (Runs# 23017 – 39963) 1059 Runs a 3 k events 2. 99 M Triggers Time 7: 05: 33: 39 Normal Physics Data-taking: • MEG 11 -mixed trigger 6. 5 Hz Trigger Rate, LT~ 80 -83% • Daily LED-calibration beam “off” • 3/week Full-calibration LED beam “on” +LED beam “off” + C-W (Li) + C-W (B) + s • 1/week 24 Hrs RMD 5 -mixed trigger data Peter-Raymond Kettle MEG Review February 2009 Total of 139 TB Data Taken 2008 17

Physics Data Preparation Analysis Scheme: (Physics Analysis Working Group) Scheme • Data Reduction in Physics Data Preparation Analysis Scheme: (Physics Analysis Working Group) Scheme • Data Reduction in form of “Pre-selection” Pre-selection - use very lose cuts “Conservative Criteria” (ensure non-biasing) reduces data to 16% of triggered events • Incorporate “Blinding” in “pre-selected” data use “Hidden” Signal-box on parameters E & Te directly via MEGAnalyzer with widths ~ ± 4. 8 Me. V & ± 1. 5 ns respectively • Perform Likelihood Analysis on “final revealed data” after optimized background study outside “signal-box” “side-bands” n tio a ul m Si “Blinding” • Probability Density Functions (PDFs) for Likelihood Analysis obtained partially direct from measurement & partially from MC. n tio a ul m Si • MC substantially advanced e. g. RMD + radiative corrections etc. Peter-Raymond Kettle MEG Review February 2009 18

Conclusions • With a consolidated effort made by the “whole” collaboration, as well as Conclusions • With a consolidated effort made by the “whole” collaboration, as well as basically starting from ”scratch” at the beginning of the 2008 we were able to achieve our goal of taking “True” Physics Data! • despite many detector/electronics problems that were encountered we were able to dedicate 12 weeks out of the expected 16 weeks to “pure (MEG+RMD) Physics Data • a vast amount of calibration data was taken during the whole 2008 period which served a as vital input to understanding encountered effects during the run – this however will continue to serve as a basis for a better understanding of our detector with on-going analysis • several factors concerning our hardware led to a non-optimal MEG Detector in 2008 the main issues have been addressed ( DC: HV-stability, Calorimeter: LXe-purity, PMT gain-stability, TC: fibre incorporation in trigger) • the most worrying issue is that of the DC HV-stability – this however is being tackled with a large and dedicated effort by the “detector group” – the experts! and as has been shown before, especially with “forefront” detector technology such problems CAN BE SOLVED! • we still have a lot of work to do & a lot of improvements are still necessary but !!! the following “Expert” talks will show that the MEG Collaboration has a lot of dedicated & resourceful means at it’s disposal Peter-Raymond Kettle MEG Review February 2009 19