ECAL DPG

Paolo Meridiani - INFN Roma1 1

ECAL DPGECAL DPG

P. Meridiani

CMS Italia

13/02/2007

http://www.roma1.infn.it/main/mappa_sito.html


OutlineOutline

ECAL Achievements in the 2006Task list & organization of the new

ECAL DPG



Achievements in the 2006Achievements in the 2006

From June 2006 ECAL community has been splitted into several experimental setups

TB @ H4 (ECAL) TB @ H2 (ECAL + HCAL) Cosmic setup MTCC

Several achievements in different technical areas Full transition to CMSSW:

• LocalReco • Simulation • DB • DQM

DAQ & Trigger firstly integrated and validated Laser operations towards final setup

Italian contributions has/will been/be flagged

P. Meridiani, A. Ghezzi

F. Cossutti, C. Rovelli

F. Cavallari, S. Rahatlou

G. Della Ricca, F. Cossutti, B. Gobbo, A. Ghezzi



H4H4

4 periods, from July 24th to Nov 12th, for a total of ~70 days of data taking, 6300 runs, ~ 2 billions electrons

To summarize: First goal:Intercalibration (9 SM (1 twice), 1/4 of EB)) Energy Scan & linearity (15-250 GeV, in M1, M3, M4 and as function

of eta) Gap & cracks, corners & edges Xtals Monitoring and Laser operation Irradiation (5 Xtals) Trigger primitives Zero suppressions Synchronous running Muons



Intercalibration at H4 [1]Intercalibration at H4 [1] 9 SM put on the beam, of which one repeated twice

Better than what expected considering that beam arrived with 1 month delay

At regime intercalibrating 1 SM (1700 xtals) took less than 2 days Different intercalibration methods tested: intercalibration using the

single crystal response

Intercalibration reproducibilityin a two months period (SM22)

5x5 resolution at 0.7% @ 120 GeV

Corrected single crystal response

All 9 SMs

A. Benaglia, P.Govoni, A. Martelli, M. Paganoni, D. Riparo, R. Salerno, V. Tancini



Intercalibration at H4 [2]Intercalibration at H4 [2]

In situ intercalibration methods: based on the energy reconstructed in a cluster (typically 5x5)

Different techniques can be used to evaluate single crystal intercalibration (matrix inversion, iterative algorithms)

Performances compatible w.r.t. single crystal intercalibration

A. Benaglia, P.Govoni, A. Martelli, M. Paganoni, D. Riparo, R. Salerno, V. Tancini



Energy/position resolution, noise

Energy/position resolution, noise

Analysis is still on going Basically a confirmation of the 2004 results, where most of

the details of the amplitude reconstruction were understood Noise in ADC countsEnergy resolution in 3x3 using 3+5 weights

Preliminary

Preliminary

C. Rovelli



More effort has been put on MC validation/tuning/comparison w.r.t 2004. Last possibility to tune MC before CMS

In particular Lateral containment

On going work together with G4 developers to understand/correct discrepancies

Tuning the MCTuning the MC

Using both G4.7 & G4.8 shower is 1% broader in sim (looking at E1/E25) Situation is reversed w.r.t G4.5 (1% narrower sim), change seems to be due to variation in the implementation of multiple scatteringRelative variation with and are instead very well reproduced

F. Cossutti, C. Rovelli, P. Meridiani



Dead channel correctionDead channel correction H4 data can also be used to evaluate dead channel corrections Main idea is to measure in the data the correlation between the energy

content of the non-functioning channel with the channels in a 5x5 crystal matrix: Correlation functions depend on the , of the crystal,particle’s energy and impact position

Simple approach and more refined (neural net) have been explored. Both seem to work fine

Energy (GeV)S25re

co_c

orr

/S25re

co

Resolu

tion

(%

)Corrections being implemented as a standard CMSSW moduleAlso map of noisy/dead channels being implemented in Offline DB



25ns bunch structure25ns bunch structure

A 25-ns bunched beam was delivered for 10 days (end of september) to H4:

Accelerator clock and orbit signals received by a LHCrx module and distributed to CCS/DCC/TCC via the TTCci.

Trigger Primitives acquired and TCC data written in the event. Used to validate CMSSW TPG emulator

The 25 ns beam sub-structure as seen from H4 TDC.



H2H2

Lots of data with different beams (e, - , - on target) with E (1300GeV)More difficult experimental condition environment with respect to H4For example ECAL temperature is not as stable as H4

Also beam contamination: delicate beam cleaning/particleId required for all analysis



ECAL intercalibration @ H2ECAL intercalibration @ H2

Intercalibration methods used at H4 have been adapted for H2 running

Temperature variations (±0.1C) (corrections needs to be applied before calibration)

Less statistics for each xtal than H4 Wider beam Beam contamination

S1 calibration results

After cal.

Before cal.

ADC counts

5x5 resolution @ 50 GeV ~ 2% Still larger than expected.

D. Del Re, R. Paramatti, S. Rahatlou



Combining ECAL+HCALCombining ECAL+HCAL

On going work: optimization of the combined ECAL+HCAL response

Choosing optimal weights for combining ECAL + HCAL energies

• E=EECAL+EHCAL

Different calibration schemas are being tried in order to optimize resolution and linearity



0 run @ H20 run @ H2

20GeV beam

0 0

Runs taken at 3 energies:9, 20, 50 GeV

Being used to optimize the pi0 reconstruction/selection

A simple test of the 0

calibration algorithm has been performed

Before calibration After calibration

Resolution using S1 calibration:(0) ~ 4.8%(0) ~ 3.7%

D. Del Re, D. Franci, S. Rahatlou



Cosmic IntercalibrationCosmic Intercalibration

Every integrated SM is/will be tested for around one week using a specific cosmic setup (located in H4)

10° inclined SM APD HV raised to have gain 200 About 5 million triggers are collected for each

SM

Useful for two main reasons Initial commissioning of the integrated SM Preliminary intercalibration coefficient can be

computed at a precision of around 2% for all channels

Systematics w.r.t to beam intercalibration are under detailed study

F. Ferri, A. Ghezzi, S. Ragazzi, T. Tabarelli



ECAL @ MTCCECAL @ MTCC

2 SMs at 11 o’clock Gain 200 (S/B 30 for aligned muons) Events with muon signal in ECAL

0.5% of DT trigger 0.15% of CSC trigger

Checks with/without magnetic field Pedestals/Noise

• Observed variation of the value of the pedestal but noise is unchanged

Muon signal amplitude/timing• Amplitude unaffected, small variation of the timing (3ns) still under investigation



DAQ & TriggerDAQ & Trigger

ECAL DAQ software system: The ECAL DAQ software system has been used in H4, H2 (HCAL+ECAL)

and MTCC (integrated with Global DAQ & Trigger) Final H4 test-beam version (release ECAL_0_1_0, based on XDAQ_V3.5.2

and SLC3) has been closed and archived. Work to port it to XDAQ_V3.7.3, SLC4 /64-bits is in progress.

In H4 also full DAQ Hardware integration has been achieved Full ECAL triplet: CCS, DCC, TCC MATACQ board: 1-ns sampling of laser pulse for monitoring purpose

TODO: Move toward SLC4 / 64-bits Software system is done to support 36 SMs in EB. To be adapted to

include Endcaps DAQ and Trigger configuration from DB (tables implemented, but read

now from XML)



DatabasesDatabases

The DB and its access tools have been operated at H4 and H2 test-beams

DCS Configuration DB (via PVSS) has been tested and works fine. DCS Condition DB (via PVSS) has performance issues. DAQ/Trigger Condition DB: run records and DAQ/Trigger

configuration keys are written in the Condition DB. Writing of summary histograms will be implemented.

DQM (Data Quality Monitor) logging to Condition DB works well. DCU measurements (APD temperatures+currents, electronics

temperatures, LV settings) are written to Condition DB via C++ API.

Most offline DB objects have been implemented and integrated into the CMSSW reconstruction

F. Cavallari



DQMDQM

ECAL DQM has been widely used in all the setups 2006 has been spent in continuous update/development/upgrade Many useful low level detector tasks have been implemented, used to

spot problems/debug the system• data integrity• pedestal mean/noise• signal pulse shape• laser/test-pulse amplitude & timing response• trigger primitives activity

Some TB physics oriented task have been also developed• Cosmics occupancy and energy spectrum• beam energy deposit (max crystal, 3x3 matrix)• beam profiles as measured by hodoscope• energy deposit vs. beam position

To be moved towards the CMS physics

F. Cossutti, G. Della Ricca, A. Ghezzi , B. Gobbo



An example of a H4 DQM page An example of a H4 DQM page

F. Cossutti, G. Della Ricca, A. Ghezzi , B. Gobbo



Laser monitoringLaser monitoring

Laser: needed to track xtal trasparency changes. ~1300 Laser runs taken in the H4 setup.

This data volume is equivalent to 10 hours of full CMS running

All data at H4/H2 have been processed ‘quasi-online’ using the DQM infrastructure. However, this is different from the online “laser farm” that we’ll need in CMS

Stability : APD/PN ratios for each channel

Overall stability good, even without refined corrections. But detailed workflow from laser events to corrections is

still to be demonstrated...



Laser MonitoringLaser Monitoring H4 stability

H2: In H2 night/day effects are larger (no room with stabilized T). Anti-correlation between T and APD/PN (~ -2%/°C) as expected.

APD/PN

Stability 1.4 ‰ from gauss fit to peak.

Long tail under study, but overall

stability is satisfactory

H2 testbeam:

Black : APD/PN, averaged over 100 channels.

Red : T/20+1



ECAL in-situ calibration: statusECAL in-situ calibration: status

Different calibration methods as described in PTDR needs to be merged into a coherent plan

Since PTDR: big effort has been put on 0 calibration 4 internal notes have been produced by different groups To exploit full potential these events need to be selected

from L1 accepted events Special stream/trigger techniques are being designed

Additional effort: improve/put calibration in CMSSW. 2007 should be dedicated to

Coherent plan for calibration including first year data taking Definition of detailed workflow

D. Del Re, D. Franci, P. Govoni, M. Paganoni, A. Palma, R. Paramatti, S. Rahatlou, R. Salerno



TasksTasks

DQM Coordination and responsible for basic services Many of the other tasks listed below involve or require DQM

Amplitude reconstruction And digitization

DCC operation Data unpacking, and Digi to Raw translator; endcap specific mapping

issues

TCC operation Trigger Primitives

Emulation

Selective Readout Processor Emulation; algorithm(s) for low luminosity

Synchronization and timing



Tasks[2]Tasks[2] Databases Calibration and alignment

Calibration methods and algorithms• Use of “precalibration”• Phi symmetry 0 • Z→ee• Z→• Single electron

“low level calibration”• Pedestals; gain ratios; dead channels

Crystal transparency corrections (Laser) Alignment

Clustering And cluster corrections

Simulation Geometry (new into CMSSW 140) Shower simulation



Tasks [3]Tasks [3]

Preshower issues Try to integrate the preshower aspects within the other categories

Event display Global performance monitoring

alias offline DQM



Sub-boxologySub-boxology ECAL framework

Data format, raw data, time frames,amplitude reconstruction, application of calibration constants

Reconstruction Clustering, superclustering and corrections

Calibration and alignment Calibration Laser transparency corrections Alignment Low level calibration (pedestals, gain ratios…) ?

Simulation Geometry Shower simulation (GEANT and FAMOS)

Databases DQM Testbeam analysis



ContactsContacts Offline: Reconstruction

Low level reconstruction: Paolo Meridiani (ad interim) Clustering: David Futyan (ad interim)

Offline: Simulation Overall: Fabio Cossutti (ad interim) Geometry: Brian Heltsey

Offline: Level-1 trigger Trigger primitives: S. Baffioni

Offline/Commissioning: calibration and alignement Overall: to be defined Alignment: Brian Heltsey Transparency correction (laser): to defined

Offline: Analysis tools Requested…

Physics: JetMet Requested…



ConclusionsConclusions

2006 has been a fundamental year for the ECAL project 10 SM have been intercalibrated with the beam All integrated SM undergo to cosmic calibration DAQ + DQM integrated and tested in different setups First validation of CMSSW Offline software + DB achieved

2007: solid base to start from but refinement/completion work is

needed in many areas Primary goals

• Prepare everything for commissioning • Deliver a full coherent picture for calibration/corrections

Italian community plays an important role inside ECAL DPG covering many aspects/tasks


Documents

ECAL DPG