Update on luminosity measurement using TileCal

1
Update on luminosity measurement using TileCal
Ulla Blumenschein IFAE Barcelona

• The Tile Calorimeter status
• The Tile MinBias status
• Systematic effects
• Gains non-linearity
• PMT non-linearity
• Impact of noise
• Strategies to improve
• sensitivity
• Summary

Ulla Blumenschein
1 9/7/2006,
ATLAS Overview Week
2
Tile Calorimeter
EBA
LBA
LBC

• Barrel part (h lt 1.7) of the ATLAS
• Hadron Calorimeter
• Sampling calorimeter iron/scintillating
• plates (tiles), perpendicular to beam
• axis
• Tiles grouped to readout units (cells)
• 73 cells per Df 0.1 arranged in pro-
• jective towers with Dh 0.1(0.2)
• -gt 4672 cells
• 3 layers A/BC/D 7.2 lhad
• each tile read-out by two PMTs
• Additional scintillators in rapidity gap
  

EBC
Ulla Blumenschein
2 9/7/2006,
ATLAS Overview Week
3
TileCal Readout
Cesium
In-situ Physics
Tile
Minimum Bias
Fibre
TileCal cell
Laser
Mixer
PMT
Various readout paths Physics data sampled
every25ns LVL1 trigger readout cell towers
Integrated data10ms (monitoring)
PMT Block
Divider
Charge Injection
3-in-1
Integrator
L H
Mother Board
Digitizer
Analog
Canbus
ADC-I
Adder
Digital
Drawer
Optical Interface
TTC
ROD
µ Trigger
Had Trigger
Energy
Ulla Blumenschein
3 9/7/2006,
ATLAS Overview Week
4
Tile MinBias
Mean energy deposition O(MeV)

• 23 inelastic p-p interactions per BC
• (700 M int./sec) at nominal lumi
• Typically low-energy forward jets,
• mostly non-perturbative processes
• (few very hard interactions -gt physics)
• Large fluctuation of energy deposition
• in a given cell
• Average MinBias signal spans a broad
• range of frequencies and amplitude
• depending on lumi/h/layer

• Slow integration of PMT current
• (10ms 110 LHC orbits 300000 BC
• 7 M inelastic interactions _at_10E34)
• expected detector readout rate
• 0.5 Hz
• sensitive Anode current range
• 12pA-1.8mA
• Will be able to monitor each
• cell/PMT channel online
• Expect to store data offline
• (COOL database/CASTOR)

most exposed to MinBias
least exposed to MinBias
Ulla Blumenschein
4 9/7/2006,
ATLAS Overview Week
5
Tile Calibration systems

• Charge injection system (CIS)
• - Inject tunable pulse into
• readout chain
• -gt calibrate readout chain
• -gt monitor linearity and stability
• Laser calibration system
• - inject laser pulses into PMT
• -gt monitor PMT linearity
• -gt study saturation recovery
• -gt timing

• Minimum Bias system
• - monitor online the hadronic
• response at cell (PMT) level
• - monitor the LHC performance
• (luminosity, beam quality)
• Cs Calibration system (offline)
• - Scan cells with Cs 137 g source
• -gt quality and uniformity of
• optical response
• -gt monitor calibration inbetween
• periods of data taking

Tile Calibration trigger sent during LHC gaps 90
of calibration bandwith dedicated to MinBias
Ulla Blumenschein
5 9/7/2006,
ATLAS Overview Week
6
Tile Integrator readout path
ATLAS cavern
USA 15
INTG_OUT
PMT
INTG_GND
16 message buffers per CAN line 250
kbps
1 ADC per module 16 ADCs per CAN line) 12
bit digitizer (0-5V)
PMT
PMT
ROB
PC
Pedestal Control
CAN Port
45(36) PMT channels per module in Barrel (EB)
Micro
Analog Bus
DAC
CAN Port
ADC
CAN Port
CAN Port
CANBus
CAN Port
ADC gains dynamic range -gt 2 mA
Ulla Blumenschein
6 9/7/2006,
ATLAS Overview Week
7
Tile MinBias DAQ
monitoring
DCS
data base
Lumi monitoring
COOL/ CASTOR
TileCAL monitoring
ROB
ROB
ROB
ROB
extracted data
PC
PC
PC
PC
TDAQ
Complete data 20 kbyte/sec
Ulla Blumenschein
7 9/7/2006,
ATLAS Overview Week
8
Tile Commissioning status

• TileCal completely assembled in the ATLAS cavern
  and
• equipped with on-detector electronics
• FE electronics precommissioned with mobile
  readout
• Servicing and final FE commissioning in progress
  
• running ¼ of a tile partition (16 modules) in
  parallel
• Several periods of successful cosmic data taking
  over the
• last year, more foreseen for summer

Barrel moving to Z0
Barrel precommissioning status display 24
Modules used for extensive performance checks
during final commissioning
Mobile Read-out
Cosmic muon in Tile
Extended Barrel assembly
Ulla Blumenschein
8 9/7/2006,
ATLAS Overview Week
9
Tile MinBias System Status

• Back-end electronics
• installed and commissioned
• FE commissioning procedures
• contain gains calibration, gains stability,
• pedestal stability, pedestal noise
• Stand-alone MinBias runs
• with several modules at a time
• ToDo Integration into TDAQ and DCS

TileCal backend
MinBias Pedestal runs
Pedestal (ADC counts)
Time s
Integrator Gains calibration
Pedestal noise per PMT channel
Commissioning Interface
PMT
Ulla Blumenschein
9 9/7/2006,
ATLAS Overview Week
10
Gains Precission

• Use a charge injection scan (100 events per
  point, 11 points per gain)
• to calibrate the 6 gains
• Check routinely for different, more trivial
  sources of failures
• (cards not switched, pedestals not in correct
  range ...)
• Fitting gains with a precission of 10-5
• Gains are stable at a level of 10-4 -10-3 at a
  time scale of 6 weeks
• under different conditions (temperature
  fluctuations, HV on/off etc.)
• Conditions are expected to be more stable at
  LHC.

Module A41 stability of gain1 and gain 6 for
each PMT channel during 6 weeks 10E-3
Relative stability of average gain 1 and 6 in 6
modules during 6weeks 10E-3
Ulla Blumenschein
10 9/7/2006,
ATLAS Overview Week
11
PMT linearity

• quality requirements linearity lt 2 up to
• peak currents of 50mA (pulses of 20 ns)
• short term drift (hours) lt 0.8 (monitored by
  laser)
• long-term drift (weeks) lt 2 (monitored by
  laser)
• Can calibrate the MinBias current range
• (10pA-1mA) using the laser system, if needed

PMT non-linearity (2-pulse technique)
1
-gt Systematic effects expected from PMT
performance ltlt 1
TileCal laser calibration system
Ulla Blumenschein
11 9/7/2006,
ATLAS Overview Week
12
Impact of Noise

• A noise of 0.5-2(2.5) ADC counts corresponds to
• gain noise nA
• 1 0.4-1.1
• 2 0.024-0.094
• 3 0.021-0.084
• 4 0.011-0.044
• 5 0.008-0.032
• 6 0.006-0.024
• Expected signal _at_10e34
• 100-700 nA
• (first two samples)

• Fraction of noisy channels 1 (PMT DC,
  light leaks, ...)
• -gt exclude cells with a noisy PMT from lumi
  monitoring
• -gt 2 of the cells which can be used for
  lumi monitoring
• So far 1 unrecoverably noisy module
• -gt exclude module from lumi-monitoring (Df
  0.1 slice)
• expect to loose 4 modules -gt 1.5 of cells
• Spread of noise contribution from LVPS use
  only PS with
• low noise 0.5-2.0(2.5) ADC counts,
  including PMT DC noise

• Very conservative approach assume a loss
• of 10 of the cells for lumi monitoring due
  to noise

Noisy module (LBA45)
Noisy channel (LBA44)
HV ON
HV OFF
Ulla Blumenschein
12 9/7/2006,
ATLAS Overview Week
13
Reducing Integrator Noise

• Recover noisy channels by Low voltage
  retrimming

LBA43 after retrimming
Low Voltage PS and DCS control panel
LBA43 before retrimming

• Prevent Voltage oszillations in Low voltage PS
  

Noise after modification, that prevent
oszillations
Noise with Voltage oszillations
Low Voltage oszillations
Ulla Blumenschein
13 9/7/2006,
ATLAS Overview Week
14
Improving statistics

• Large variation of MinBias energy deposition
  with shower depth and h
• Online MinBias monitoring of all cells (2-2.5sec
  for all cells)
• For lumi monitoring
• dedicate more (all) bandwidth to the cells
  with large energy deposition
• (1920/4670 cells, mostly A-layer)
• Skip noisy cells for lumi monitoring (-gt ca
  1700 cells for lumi monitoring)
• -gt increase data taking rate up to a factor of
  2.5 complete readout with 1 Hz

Anode current nA
------------------------------------ --- MinBias
monitoring --- -----------------------------------
-
--- Lumi monitoring ---
Expected MinBias current vs eta and layer

• Expected mean current _at_10E34 250-750 nA, rms
  4nA (1) ? 1 second
• Expected noise _at_10E34 (highest gain) 0.4-1.1 nA
  (ltlt1) ? 1 second

• Average over 1700 cells Statistical precission
  of 0.02 per second
• -gt At 10E34 limited by systematic effects from
  PMT non-linearity
• (hopefully calibratable with laser system)

Ulla Blumenschein
14 9/7/2006,
ATLAS Overview Week
15
Summary

• The advanced state of TileCal commissioning
  allows for dedicated
• checks of the precission of the Minimum Bias
  readout
• Systematic effects are small
• stability of gains lt 0.1
• non-linearity of PMT ltlt1
• In addition the stability and linearity is
  monitored
• online and offline by dedicated calibration
  systems.
• Noise contribution (external noise, power
  supplies, PMTs)
• 0.5-2.5 ADC counts in gt 90 of all cells
• ? rms/signal ltlt 1 /sec at nominal
  luminosity in cells
• to be used for lumi monitoring
• The statistical uncertainty can be reduced by
  dedicating more (all)
• bandwidth to cells with large energy
  deposition and low noise
• -gt statistical precission opf 0.02 per
  second at nominal luminosity

Ulla Blumenschein
16 9/7/2006,
ATLAS Overview Week
16
ATLAS startup
LHC Startup 900 GeV
LHC Final 14 TeV

• LHC startup with reduced beam energy
• Ecms 900 GeV
• ? expect reduction of average MinBias current
  by
• factor of 5
• Lower startup luminosity (? 10e29 )

Expected Performance , average 1700 cells per
second/hour
Lumi MinBias intrinsic max
noise rms rms (0.9 TeV) current nA
rms nA nA (sec) (hour)
510e34 350 0.085 0.027
0.02 0.0004 510e32 3.5
0.0085 0.0006 0.2 0.004
510e30 0.035 0.0009 0.0006
3 0.05 510e28 0.0004
0.0001 0.0006 150 2.5
Expected MinBias cross section, particle
multiplicity and mean trans-verse energy as a
function of vs
Ulla Blumenschein
15 9/7/2006,
ATLAS Overview Week