Calo preparation for 2015

Calo preparation for 2015

Goals:-Trigger stability-Good calibration for HLT2 processing-Improved calibration ( timing, e/gamma response) for all calo sub-detectors-Improved monitoring

Calibration workshop 8/10/2013• The hardware changes :

– Damaged Ecal LED fibers will be changed -> improved gain monitoring (Iouri)– Exposed PMT will be replaced ( hundreds)

• The current calibration scheme for ECAL/PRS– Initial calibration from pi0 fine calibration (D.Savrina) using FMDST available ~ 4 weeks of data taking

• Minimum biais events . Low energy in PS • ~ 80M events needed• Iterative procedure • The more precise method

– Ageing effect corrected from e/p of tagged conversions 55 sets in 2013 (MNM) – Calibration from electron provide also calibration constant also for pre-shower. (O. Steinakin)

available only after reconstruction. • Analytic solution

• The current calibration scheme for Hcal – Ageing monitoring through LED response (2012)– PMT change when deviation above certain threshold – Controlled by Cs source run ( every TS)– Fine Hcal calibration on reconstructed data :100-200M events needed to get a cell/cell calibration.

Yu. Guz 2013/10/09 3

LED system: HCAL 2012

sum/sumref, 56 central cells

shut

dow

n 13

7 Cs c

alib

ratio

n

TS1,

137 C

s cal

ibra

tion

TS2,

137 C

s cal

ibra

tion

TS3,

137 C

s cal

ibra

tion

LED-based HV corrections (17) Cs calibrations (3)

results of Cs calibrations at TS is used as a starting point, then LED-based correctionstotal of 4 Cs calibrations and 17 LED-based corrections

Precision of LED corrections is limited by: annealing during TS (and faster ageing afterwards) - PMT rate effect – variation within fill (can be optimized by regrouping PMTs, to be done @ LS1)uncertainty in the “plastic ageing” prediction – non linearity, annealing

(the attempts to take the plastic ageing into account are visible)

Occupancy Method & dependence• For each cell calculate the occupancy ratio :

- Threshold to be determined – Precision to be reached 1-2%Irina Machkhilyan LHCb-INT-2013-050Calo meeting presentations from Irina in each calo meeting since July 2012.http://indico.cern.ch/contributionDisplay.py?contribId=9&confId=217410http://indico.cern.ch/contributionDisplay.py?contribId=8&confId=235526http://indico.cern.ch/contributionDisplay.py?contribId=7&confId=244405http://indico.cern.ch/contributionDisplay.py?contribId=5&confId=252503http://indico.cern.ch/contributionDisplay.py?contribId=6&confId=260472

Occupancy ratio #fill dependence

MagDown Magup

- Depends upon the fill m Stable period in m to be used

-Depends of the trigger type -> Minimum biais 85 Hz for 8 hours fill enough to reach a 1% precision

- Magup/Magdown reference

Can be extended to HcalHcal / LED response

Ecal/LED correlated to HV change

PRS case

Determination of the integration range complicated By the scale change of the adc spectrum .

Checked on data as well.

Method’s checks on 2012 data

* Several run used to calculate the correction from occupancy . Bothpolarity

Reference are in italic

Ecal

• E/p fit

Hcal

Removing the Hcal HV change effect calculated from LED

On raw Hcal fluctuation at 2% level .More investigation on PMTs affecting the triggerThe accuracy of the method can be compared to the LED

PS

Ecal corrected

Ecal +PS corrected

E/p for electrons

-Ecal occupancy correction applied -Ecal & PS occupancy correction applied

-Overcorrection when both are applied:

-To be understood

-Impact on L0 to be investigated but is expected to be small.

For 2015 • 85 Hz minimum biais needed ~ 300K /hour

– 1 :1.5 % precision all over the calorimeter (10 hours)– 2 consecutive fill can be used– Choose stable m period– Raw information should be available for end of fill processing

• Built tools able to determine runs with stable mu • 2 options

– Correction table / cell – HV change – preferred will solve the trigger instability

• Cannot be reversed• PMT HV curves used

• With new LED system the ageing amplitude can be monitored and used to cross check occupancy changes.

• Initial calibration needs other methods – P0 (Ecal)– Electrons (Ecal +PS)

Is it possible ?• Need 85Hz of minimum biais event to calculate occupancy cell by cell

( occupancy farm ?)• Analyse events only in mu stable condition • Reference depends upon the magnetic field• The PMT HV can be updated when the deviation > 2-sigma of its resolution• Possible to check the agreement with LED measurement (Ecal & Hcal)

• This method had been used in November 2012 to update Ecal HV• It will remove the ageing dependance from trigger• Is it possible to fit with the new HLT1/HLT2 scheme• It has to be completed by fine calibration

• To be checked in event replaying mode . ( 1st 2014 half)

Data quality controls• Improved data quality to point wrong behavior:

– On line (in investigation) :• The possibility to use the 85 Hz mimimum biais event to check the pi0 mass

reconstruction as done in the fine calibration: with or without iterative procedure– e/p per zone for conversion to become standard plots– Pi0 Mass /zone will also become standard DQ plots– e/p on hadronic charged tracks

• Current developement ( available in 2014)– Retuning of all corrections for electrons and pions– Clustering (2x2 investigate for the upgrade )

• Initial precise calibration will come from pi0 , electrons , hadrons: FMDST scheme still needed.

spares

2012 procedure• Reference calibration from reconstructed p0

– Processing using the current calibration– Extract correction coefficient (Dasha)

• Need FMDST to be available • Processing 2 weeks

•

•

All 1

Reference used

Ageing trend general

Lumi (pb-1)

The correction are dominated by the higher flux in each zone.They correspond to higher ageing.

HV change end of august~1% /100pb-1

HV change: occupancy methodEnd of November .Back to initial calibration.~0.4%/100pb-1

Annealing effect : Christmas stop4% annealing in average

Frac

tion

gain

lost

E/p for electron (O.Steniakin)

Fine calibration applied on already calibrated data

Yu. Guz 2013/10/09 19

HCAL E/p calibration: track selectionFor all tracks from Rec/Track/Best : find associated cells in HCAL, ECAL, PS laying at the track extrapolation; calculate E(HCAL), E(ECAL), E(PS)

using the CaloEnergyForTrack tool (Vanya) (AddNeighbors=1) keep only tracks with E(HCAL)>0 require isolation in HCAL (no HCAL cells associated with given track is shared with any other track); keep only long tracks, -200<z<400mm reject muons (tight and loose) reject protons (by RICH): the HCAL responses to and differ, which distorts calibration require E(ECAL) < 800 MeV && 0< E(PRS) < 20, to reject showers started much before HCALThe track position in HCAL is defined as the position of a cell with max energy deposition. further refinement:

cut on ptrack: 7 GeV for all cells, somewhat higher for the innermost ones (safety) cut on E(HCAL): 2.5 GeV for all cells, somewhat higher for the innermost ones

p p

Yu. Guz 2013/10/09 20

HCAL response to hadrons, data

More uniform than response to muons: ~4% RMS for both inner and outer (borders excluded).