Curious Facts about CMS HCAL

1
Curious Facts about CMS HCAL

• Jim Freeman
• Fermilab

2
CMS Detector
HF
HO
HB
HE
3
Hadronic Calorimeter HCAL
Had Barrel HB Had EndcapsHE Had Forward HF
HO
HB
HF
HE
4
CMS at LHC Point 5
UX5 about 100 meters underground
5
Civ Eng SX5 and pit-head cover

6
Experiment Cavern UXC55 delivered to CMS
UXC55 Feb 05
7
CMS in UX5
8
Heavy Lowering
Heavy lowering starts May 2006. 15 major lifts.
9
USC55 Relay Racks
USC55 Beneficial Occupancy (i.e. can install our
crates) Mar 2006
10
HCAL-Absorbers
HB
HE
HF
11
HCAL HB Calorimeter (Barrel)
Sampling calorimeter brass (passive)
scintillator (active) Coverage
hlt1.3 Depth 5.8 lint (at h0)
segmentation f x h p resolution
120 / 0.087x0.087
17 layers longitudinally, f
x h 4 x 16 towers
Completed assembled
20o
f

12
HCAL HE Calorimeter (Endcap)
Sampling calorimeter brass (passive)
scintillator (active) Coverage
1.3lthlt3 Depth 10 lint
segmentation f x h p resolution
120/ 0.087x0.087



19 layers longitudinally
20o
Completed, assembled, HE-1 installed
13
Reading out the Towers of Tiles with WLS fiber
and HPDs
14
Optics-Megatiles

• Cross section view of a megatile

15
Optics-Megatiles

Components are the machined scintillator plates,
cover plates, fiber assembly (WLS spliced to
clear fiber, optical connector) pigtails
16
Drew, what is he doing?
17
Installing Megatile
18
Scint brightening vs B
Scintillator emits more light as function of B.
(Endcap configuration)
19
Longi brightening
Perpendicular (Endcap)
Parallel (Barrel)
Longitudinal shower development vs magnetic field
when field is perpendular ( parallel ) to
absorber plates
20
HCAL and e/? in B Fields
21
HCAL - Path Length/Loopers
22
Installing Megatile
Phosphor-bronze spring to hold megatile at large
radius. Minimize Bfield effect
23
Readout Module Overview

24
Optical Decoder Unit
HPD mount aligned to cookie and plate
Optical decoding from layer into tower bundles
occurs at the readout boxes.
Fiber Optic Cables attach to a patch panel
25
RBX Readout Module

• The readout module (RM) integrates the HPD, front
  end electronics, and digital optical drivers.

26
6 Channel FE Board
AD590 Temp Sensor
Low Voltage Regulator (2 / board) CERN Developed
in rad hard process
P82B96 I2C Transceiver (change to P82B715)
Custom ASICs
QIE (6 / board) Fermilab
MC100LVELT23 LVPECL-LVTTL
Honeywell VCSEL HFE419x-521 (2/board) back side
of board
MC100LVEP111 LVPECL clock fanout chip
Gigabit Optical Link GOL (2 / board)
CERN Developed in rad hard process
CCA (3 / board) Fermilab
OP184 bi-polar OpAmp
PZT222A transistor
27
QIE Flash ADC

Implementation of a 4-range splitter and a 5 bit
non-linear FADC allows to simplify the design
while having minimum effect on the detector
resolution.
Green is multiplicative increase in measuring
error in
28
QIE 4 ranges, 32 bins/range

• Dynamic Range (GeV)
• HB/HE
• 15 pe/GeV
• 2000X HPD Gain
• 2 TeV Saturation per time slice
• HF
• 0.25 pe/GeV
• 105 Gain
• (27 fC full scale)
• 8TeV Saturation (nominal)
• Because of pmt variation, 2.7 TeV guaranteed

Note because of nonlinear binning, plotting raw
distributions very misleading/meaningless
29
HB RBX Assembly
Full RBX with 19 ch RMs
RBX Interior -- HV distributor and backplane
30
HCAL FE/DAQ Overview
SLINK-64
READ-OUT Crate (in UXA)
Trigger Primitives
Rack CPU
VME
D C C
D C C
12 HTR, 2 DCC per crate
C L K
FRONT-END RBX Readout Box (On detector)
HPD
Shield Wall
Fibers at 1.6 Gb/s 3 QIE-channels per fiber
FE MODULE
31
HPD a long, expensive battle

Diode Structure
Bias Voltage
32
Capacitive Crosstalk eliminated by Aluminization
Eliminating capacitive cross-talk revealed
another type of cross talk (optical)
High resistivity diode connection to Vbias caused
capacitive cross talk. Needed layer of Al
33
Test Confirms Reflected Light
Light injected thru fiber
APD views reflected light
Light
Re-emitted photoelectrons
photoelectrons
pe backscatter
34
HPD Anti-reflective coating

Aluminum with no anti-reflective coating
Reflectivity vs wavelength for diode with
antireflective coating
35
Yet another problem/feature Electron
Backscattering off HPD Diode
36
e- Backscatter crosstalk
Move fiber
Read outnearest neighbor
d
Convolution of hexagonal pixel shape with
backscatter radial distribution
37
e- Backscattering focused by B field
Inject light onto one pixel, observe signal on
nearest neighbor as function of B. Note
saturation at B 0.3T. So B field eliminates
problem. Note that B causes image shift so need
to align HPD E and B to 5o. Also note that
this effect needs to be corrected for when
transferring calibration from testbeam to CMS
38
Measure B direction using sharing
HO ODU has special calibration fibers between
pixels. Can measure change in sharing vs B.
Aligned ? no change.
39
Shorten pulse width by thinning diode
Drift time is approximately given by and the
shape of the plateau mirrors the internal
electric field
300 mm thick
200 mm thick
40
HPD Diode Development Saga
2 side-contacts (100 nm thick Al)
Bump-bonded vacuum feedthru
41
Radiation Damage to HPD? increase in leakage
current
Integrated dose is geometrically equivalent to
4.5 x 1011 n/cm2 head on
Flux 7 x 109 n/(cm2 hr)MeV neutrons from Cf252
HPD (light-dark)
Note initial increase which is characteristic of
HPDs
PIN diode gt Injected light is constant
Leakage current rising at 46 pA/hr
hours
10 CMS years
Day 1
Day 2
Day 3
Day 4
42
HB Pulse
Jordan Damgov
43
HB Pulse Shape (25ns bins)
100 GeV pions
100 GeV electrons
Each histo is phase-shifted by 1ns relative to
trigger,modulo 25ns
44
Correlation between Phase time and Calculated
time (using event shape)
HE
Calculated event time (in clock cycles)
Measured clock phase P
Pulse shape very reproducible. ? Can use
theoretical pulse shape to compensate if we ever
see saturated pulses.
45
HB Charge Collection 1 bucket vs 2 pions and
electrons
Pions and Electrons have very similar pulse shape
and relative timing
46
Measuring relative phase of different towers with
laser
Laser calib system designed to mimic beam timing.
It works. Laser and pion measurements track. ?
Can use laser system to measure relative phases
of different towers and then program CCA to
compensate.
TIME ?
Eta Tower number ?
47
Relative Phase after alignment
Use laser to calculate d(t) offset. Program CCAs.
Use electron beam to check. ? Can synchronize to
lt2ns using laser system
48
Setting timing of both wedges simultaneously
Use universal phase constants to set both
wedges. Conclude Rms(w1 or w1) 1.2ns
d(w1-w2) lt0.2ns Wedges are identical. Can use
laser to predict overall timing at CMS up to a
constant
Wedge 1
Wedge 2
N towers
( Tower phase - mean ) (ns)
49
Interesting Observation about QIE
Medium sharing
Low Sharing
Note apparent phase shift of data as function of
energy deposition in tower. ? QIE impedance
change vs input current level Needs corrective
action. Will optimize input slew vs noise level.
Medium sharing
Low Sharing
Low Sharing
50
Time Slew vs energy
1
Noise
slow
Previous plot
medium
2
Final choice (except HO)
3
fast
1fc 6000 e
1fc 1 ADC count
0
ideal(noise 3800 e)
1fc 250 MeV
5 GeV
50 GeV
500 GeV
51
HCAL HF Calorimeter (Forward)
Steel absorbers, embedded quartz fibres // to the
beam Fast (10 ns) collection of Cherenkov
radiation. Coverage 3lthlt5
segmentation f x h Depth
10 lint 10o x 13 h towers
52
HF Fiber Spacing

• 0.6 mm quartz fibers in iron
• Half a million quartz fibers viewed with about
  2000 phototubes

53
HF
HF are first Items to be lowered in May 2006
Both ends assembled in Bat 186
Individual wedges
54
HF Pulse Shape (2ns bins)
25 ns
55
New Projects and RD

• HO Trigger
• Replacement for HE high eta scintillator (quartz
  plate/wls fiber)
• New readout for HO? (SPMT)
• Improved calorimeter Trigger
• Luminosity Monitor using HF
• Implement/verify HCAL sim in CMSSW
• Better instrumentation of 06 testbeam

56
Pedestal dist HO channels

• Rms per channel
• ltgt 1.6 QIE counts calib mode
• 3200 e-

57
SLHC Detector Environment
Bunch spacing reduced 2x. Interactions/crossing
increased 5 x. Pileup noise increased by 2.2x if
crossings are time resolvable.
58
Scintillator - Dose/Damage
Current operational limit 5 Mrad
59
Radiation damage to scintillators

Dose per year at SLHV
ECAL HCAL
5 Mr
Barrel doses are not a problem. For the endcaps a
technology change may be needed for 2 lt y lt 3
for the CMS HCAL.
60
Quartz Plate Simulation Geometry

• 10 x 10x 0.2 cm Quartz plate
• 0.35/0.30 mm radius clad/core wls fiber
• Counting the photons (wls)
• reaching the PMT
• PMT photo-cathode detection.

61
Quartz plate/WLS Readout (Iowa)

• We tested UVT and Quartz plates, UV absorbing
  wavelength shifting fibers and liquids. Also,
  wrapping materials Tyvek and Al-Mylar .
• July04 _at_ CERN
• Aug04 _at_ Fermilab
• Jan 05 _at_ Fermilab
• Cerenkov light collected from quartz plates with
  WLS fibers reached 25 of regular HE plates.
  Low-OH Polymicro solarization quartz is our
  choice over high-OH Polymicro. Peace fiber
  geometry collected light more efficiently. Small
  plate size is better.

62
Geiger-mode silicon pmt (SPMT)
40 mm
100 10000 SPMTs tied in parallel to same
substrate. High gain, 106, low noise Issues are
rad-hardness, noise, and rate ability
pes
63
Schedule

• Magnet test Feb 06
• Crates go into UX April 06
• Lowering HCAL June 06
• UX commissioning summer/fall 06
• Testbeam in summer 06
• We have a VERY busy year ahead of us!